Powder Technology 294 (2016) 55–65

Contents lists available at ScienceDirect

Powder Technology

journal homepage: www.elsevier.com/locate/powtec

Effect of flotation froth properties on froth rheology

Chao Li ⁎, Kym Runge, Fengnian Shi, Saeed Farrokhpay
Julius Kruttschnitt Mineral Research Centre, The University of Queensland, 40 Isles Road, Indooroopilly, Queensland 4068, Australia

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

Article history: Froth flotation is a widely used process of particle separation exploiting differences in surface properties. The
Received 14 December 2015 froth performance in a flotation cell is expected to be affected by the froth rheology, as it affects the froth resi-
Received in revised form 8 February 2016 dence time that determines the probability of recovery of valuable minerals in the froth phase. Flotation froths
Accepted 10 February 2016
have a similar structure to the gas–liquid foams whose rheology has been widely studied. However, to date,
Available online 13 February 2016
very little work has been done in the rheology of flotation froths owing to their instability and the presence of
Keywords:
solid particles (on bubble surfaces and in the plateau borders) that are believed to influence froth rheology
Froth rheology and complicate any investigation. In this paper, the effects of froth properties on froth rheology were studied
Bubble size by examining the results of 33 flotation tests performed under various conditions that resulted in changes in
Fraction of lamellae covered by solids the froth properties and, consequently, the froth rheology. The experiments were performed in a 20 L continuous
Flotation flotation cell. It was found that the bubble size and the fraction of lamellae covered by solids defined the froth
rheology, while the presence of particles in the plateau borders contributed very little to the froth rheology. A
model structure was developed by taking into account froth properties to predict froth viscosity.
© 2016 Elsevier B.V. All rights reserved.

1. Introduction when subject to different shear rates. In general, a substance can either
exhibit Newtonian or non-Newtonian behaviour, with the latter in-
In flotation, the froth phase plays the role of transporting the hydro- cluding dilatant, plastic, pseudo-plastic and Bingham behaviours [8,9].
phobic minerals from the collection zone to the concentrate launder. Various types of rheograms are illustrated in Fig. 1. Viscosity as a key
Froth transportation consists of both vertical and horizontal motion: rheological term is a measure of the resistance of a material to deforma-
the vertical motion is defined by the flow of bubbles that carry particles tion. It is a constant in Newtonian flow but shear rate dependent in non-
moving from the pulp–froth interface to the launder lip level, while the Newtonian flow.
horizontal motion describes the motion towards the overflow weir [1]. In order to fully understand what froth properties determine the
Drainage of valuable minerals occurs during froth transportation owing froth rheology, it is firstly necessary to gain insight into the froth charac-
to bubble bursting and bubble coalescence. Froth residence time, which teristics. From the study of aqueous foams, when the gas volume frac-
determines the probability of recovery of valuable minerals, is a func- tion is less than 0.73, bubbles disperse in the liquid phase without
tion of the time bubbles take to move both vertically and horizontally. becoming attached to one another; at a gas volume fraction greater
The vertical motion is driven by the superficial gas velocity. The hori- than 0.73, the bubbles start to pack and are separated by thin-plane-
zontal flow is a consequence of three factors that include the force of parallel films forming polyhedral cells (lamellae) [10]. The thin lamellae
gravity, the froth stability and the resistance to froth flow. The influence meet in lines (plateau borders) and the lines meet at vertices [11,12].
of the froth stability on the efficiency of froths in recovering valuable Flotation froth has a similar structure to dry foam. The air volume frac-
minerals has been extensively studied [2–5]. The resistance to froth tion in flotation froth usually exceeds 0.90 especially in deep froth due
flow is supposed to be directly associated with froth rheology, and, to rapid drainage. Furthermore, flotation froth is a gas–liquid–solid re-
therefore, the importance of froth rheology on flotation performance gime; there are solid particles present in the froth phase. Hydrophobic
has also been recognised. Shi and Zheng [6] and Farrokhpay [7] have particles are mainly attached to the lamellae while both hydrophilic
clearly shown that froth rheology can affect the froth mobility as and hydrophobic particles (detached owing to bubble coalescence and
well as the froth stability, and ultimately influence the flotation per- bubble bursting) are present in the plateau borders and vertices. A typ-
formance. However, to date, it is not clear what froth properties af- ical froth structure with particles is shown in Fig. 2.
fect froth rheology. Flotation froths have a similar structure to foams, making it possible
Rheology is a measure of the flow characteristics of a substance. It is to begin the study of froth rheology by considering the rheology of
usually represented by a rheogram which plots the shear stress of a fluid foams (i.e. leaving aside the presence of solid particles). Practically,
foam rheology is associated with bubble size and foam quality (the vol-
⁎ Corresponding author. ume fraction of air in the foam) [13–16]. When foam is dry, its rheology
E-mail address: c.li7@uq.edu.au (C. Li). is dominated by the bubble size [15,17]. However, a flotation froth

http://dx.doi.org/10.1016/j.powtec.2016.02.018
0032-5910/© 2016 Elsevier B.V. All rights reserved.
56 C. Li et al. / Powder Technology 294 (2016) 55–65

in the plateau borders and vertices form the local solid–liquid suspen-
sion. According to the study of suspension rheology, the solids volume
fraction in the suspension influences the rheology [19]. Hence, whether
the solids volume fraction in the plateau borders and vertices is also a
crucial factor affecting froth rheology needs to be investigated.
Previously, the authors have investigated the effects of various flota-
tion conditions (viz. froth height, gas rate, impeller speed, feed particle
size and feed grade) on froth rheology. A Central Composite Rotatable
Design (CCRD) study was carried out, in which 33 flotation tests were
performed in a 20 L continuously-operated flotation cell [20]. The froth
rheology was measured using a method developed recently by the au-
thors [21]. Ultimately, the flotation conditions influence the froth rheolo-
gy through their effect on the froth properties. In this current work, the
results of the previous experiments (from the CCRD study) are examined
to fundamentally investigate the direct effect of the froth properties on
froth rheology. The froth properties of interest were identified above as
bubble size, fraction of lamellae covered by solids and the solids volume
fraction in the plateau borders and vertices. The method of evaluating
Fig. 1. Schematic diagram of shear rate as a function of shear stress for different types of fluid.
these froth properties is developed below.
After Mewis and Wagner [8].

2. Experimental
differs from a two phase foam in that there are also solid particles
present — both as attached particles on the lamella of the bubble surface 2.1. Flotation tests
and in the plateau borders which form between the bubbles. It is not yet
clear how the presence of solid particles affects the rheology of flotation The 33 flotation tests were performed in a bottom driven 20 L flotation
froths. The hydrophobic particles attached on the lamellae in flotation cell with cross sectional dimensions of 30 by 30 cm. Table 1 shows the de-
froth function as the surfactant adsorbed on the interface film in aque- tails of the 33 tests in which there are seven repeat tests (i.e. Tests 1, 9, 10,
ous foam. The surfactant adsorbed on the interface film lowers the inter- 12, 18, 19, and 26). The aim of this work is to create flotation froths with
face tension and resists bubble coalescence, stabilizing the foam. The different froth properties and evaluate the effects of these froth properties
presence of surfactant on the film is characterized as its adsorbing thick- on the froth rheology.
ness [18]. In this study, the presence of solid particles attached to the la- The flotation feed was a mixture of pure chalcopyrite and silica. The
mellae can be represented by the fraction of lamellae covered by solid chalcopyrite was purchased from Geo Discoveries as bulk rock. The
particles which takes into account the mass of solid particles attached
to the lamella per unit area and the size distribution of the solid parti- Table 1
cles. It is expected that the particles on the lamellae change the bubble The conditions in the CCRD flotation tests [19].
rigidity and smoothness. In addition, the particles trapped unselectively
Test Froth Superficial Impeller Chalcopyrite Copper
height gas velocity speed particle size grade (%)
(cm) (cm/s) (rpm) P80 (μm)

1 7 1.4 900 80 1.0

2 6 1.0 750 50 1.4
3 6 1.8 1050 50 1.4
4 8 1.8 750 50 1.4
5 6 1.0 750 110 0.6
6 6 1.8 750 50 0.6
7 6 1.0 1050 50 0.6
8 6 1.8 1050 110 0.6
9 7 1.4 900 80 1.0
10 7 1.4 900 80 1.0
11 8 1.8 750 110 0.6
12 7 1.4 900 80 1.0
13 8 1.0 1050 110 0.6
14 8 1.8 1050 110 1.4
15 8 1.0 750 110 1.4
16 6 1.8 750 110 1.4
17 8 1.8 1050 50 0.6
18 7 1.4 900 80 1.0
19 7 1.4 900 80 1.0
20 8 1.0 750 50 0.6
21 8 1.0 1050 50 1.4
22 6 1.0 1050 110 1.4
23 7 1.4 900 140 1.0
24 5 1.4 900 80 1.0
25 7 0.6 900 80 1.0
26 7 1.4 900 80 1.0
27 7 1.4 1200 80 1.0
28 7 1.4 900 80 1.8
29 7 1.4 600 80 1.0
30 7 2.2 900 80 1.0
31 9 1.4 900 80 1.0
32 7 1.4 900 80 0.2
Fig. 2. Schematic of flotation froth ( hydrophilic particle; hydrophobic particle).
33 7 1.4 900 20 1.0
After Ventura-Medina and Cilliers [10].
 C. Li et al. / Powder Technology 294 (2016) 55–65 57

silica was purchased from Sibelco Australia as fine particles (P80 = size and the froth velocity profile towards the cell launder lip. A single
73 μm). Before each flotation test, a measured quantity of the chalcopy- light source was mounted above the froth surface as this results in a
rite was ground to the targeted particle size distribution, mixed with sil- single bright light on each bubble — a requirement of the froth analysis
ica to achieve the desired feed grade, and diluted with Brisbane tap algorithm. An Anglo-Platinum bubble sizer was used to measure the
water in a conditioning tank. The solids concentration was maintained bubble size in the pulp at the end of each test.
at 40 wt.% in all the tests. Sodium ethyl xanthate (2.0 g/t) and Dowfroth
250 (14.7 ppm) were used as the collector and the frother, respectively. 3. Estimation of froth properties
The flotation tests were operated continuously in a closed circuit by
recycling the concentrate and tailing. Samples of the feed, concentrate This section introduces the methods of estimating a number of froth
and tailings were collected, and weighed before and after drying. properties, viz. gas hold-up, fraction of lamellae covered by solids and
Sub-samples were assayed for copper to determine the flotation recov- the solids volume fraction in the plateau borders and vertices. To per-
ery. The concentrate flow rates in terms of mass and volume were also form the estimations, some assumptions were made based on reasoning
measured. provided in the literature.
A pictorial diagram of the experimental set-up is shown in Fig. 3.
More details of the experiments may be found in previous work [20]. 3.1. Gas hold-up (gas volume fraction)

2.2. Froth rheology measurements The estimation of the fraction of lamellae covered by solids and the
solids volume fraction in the plateau borders needs the value of the
The froth rheology measurements were conducted using a 6-bladed gas hold-up in the froth. Researchers [1,23] have found that water and
vane (22 mm diameter and 16 mm height) attached to an air-bearing particles drop back into the pulp phase mainly just above the froth–
rheometer (Anton Paar DSR301). A tube (74 mm diameter and 150 mm pulp interface, above which the froth phase is relatively constant in
height) was used to encircle the vane to eliminate the effect of the hori- terms of its properties. Hence, the mean gas holdup in the froth above
zontal froth flow [21]. The vane was positioned in the middle of the cell the lip level remains relatively constant.
with its upper edge immersed 2 cm into the froth. During the froth rheol- In a continuous and stable flotation process, the froth phase is in dy-
ogy measurement, the torque values were measured by evenly increasing namic equilibrium. Air bubbles, particles and water continuously enter
the vane speed from 1 rpm to 15 rpm in equal increments (i.e. 1.0 rpm, the froth phase from the pulp–froth interface and flow over the launder
4.5 rpm, 8.0 rpm, 11.5 rpm and 15.0 rpm), with a 5 s interval between lip. Some air is lost from the froth surface during its transportation to the
each measuring point. A total of five torque values were measured in launder lip (Fig. 4). By assuming that the drainage in the upper froth
each test. Each series of torque measurements was replicated five times phase is negligible, the gas hold-up in the upper froth zone can be
to determine the average value. The variation of the torque measurement regarded as the ratio of the air flow rate to the total froth flow rate (in-
at each speed was determined and found to be reasonably low [20]. cluding air, water and solids). The froth flow rate consists of the air flow
The vane was only immersed in the froth for the period of the rheology rate (VA) and the concentrate slurry flow rate (Vcs). As the gas entrained
measurements and then moved away not to impede the flotation froth into the tailings is negligible [24], the air flow rate is equal to the air aer-
movement. ated into the flotation cell. The air flow rate and the concentrate slurry
flow rate were both measured; hence, the mean gas holdup (εf) in the
2.3. Froth and pulp bubble characterization upper froth zone can be calculated by Eq. (1).

A digital video camera (Sony ACC-FV50B) was mounted above the

flotation cell to record the froth movement. The video images were VA
εf ¼ ð1Þ
analysed by a contracted software [22] to determine the froth bubble V A þ V cs

Fig. 3. A pictorial diagram of the experimental set up [19].

58 C. Li et al. / Powder Technology 294 (2016) 55–65

the concentrate is roughly equal to the mass flow rate of chalcopy-

rite from the pulp to the froth.

Rc  R f
R¼  100 ð2Þ
Rc  R f þ ð1−Rc Þ

The bubble surface area flow rate entering the froth phase can be de-
termined by using the bubble surface area flux (Sb) which is a measure
of the fresh bubble surface area per cross-sectional area being supplied
to the cell. It is determined by the superficial gas velocity (Jg) and Sauter
mean bubble diameter (D32) as shown in Eq. (3) [27]. The cell cross-
sectional area (A) is known. Hence the bubble surface area flow rate in
the pulp (φp) can be calculated by Eq. (4).

6J g
Sb ¼ ð3Þ
Fig. 4. The schematic of froth flow. D32

φp ¼ Sb  A ð4Þ
3.2. Fraction of lamellae covered by solids
The mass flow rate of chalcopyrite in the concentrate can be calculated
To calculate the fraction of lamellae covered by solids, the mass of from the measured mass flow rate of solids in the concentrate (∅c) and
solid particles attached to the lamellae per unit area (i.e. bubble the copper concentrate grade (G). It is known that the recovery of hydro-
loading) and the size distribution of particles needs to be known. phobic minerals from the pulp phase is mainly by true flotation, which
The bubble loading in the froth phase can be derived from the bub- means that the collection of hydrophobic minerals by entrainment in
ble loading in the pulp phase. The estimation of bubble loading in the pulp phase is limited. Thus, the bubble loading in froth phase (LB)
this work is based on the assumption that particles that detached can be calculated from Eq. (5), where ‘2.89’ is the ratio of the molecular
through bubble bursting and bubble coalescence do not reattach weight of chalcopyrite to the atomic weight of copper. As mentioned be-
on the lamellae in the froth phase. This assumption can be sup- fore, the bubble loading in the froth phase is double that in the pulp phase,
ported by some previous studies. Vera, Franzidis and Manlapig which is reflected by the number ‘2’ in Eq. (5). It should be pointed out
[25] reported that froth recovery was unselective; thus, detached that this method may underestimate the absolute bubble loading (as
chalcopyrite particles stay in the plateau borders and vertices with
no reattachment occurring. This is consistent with the finding re-
ported by Ventura-Medina and Cilliers [11], who concluded that Table 2
the recovery of hydrophobic minerals in flotation concentrate main- Data required for calculation.
ly came from the plateau borders and vertices because of bubble Jg D32 vf h ∅c Vw P50
Test R (%) G (%)
bursting and bubble coalescence, indicating that reattachment was (cm/s) (cm) (cm/s) (cm) (g/s) (cm3/s) (μm)
not significant. 1 93.73 1.4 0.043 5.54 2.30 14.82 8.15 56.03 10.21
Silica, a hydrophilic mineral, does not attach on bubble surfaces 2 89.90 1.0 0.048 3.10 2.40 7.63 11.67 28.75 8.44
and is recovered into the froth phase by entrainment, which con- 3 98.30 1.8 0.053 4.87 2.80 19.82 4.56 59.06 10.64
tributes negligibly to bubble loading. Hence, bubble loading is asso- 4 95.51 1.8 0.051 4.95 2.50 11.45 9.25 36.33 8.91
5 93.83 1.0 0.048 3.93 1.40 10.47 3.42 32.21 5.54
ciated with hydrophobic chalcopyrite particles only. In the pulp
6 93.56 1.8 0.051 5.31 2.20 14.05 2.08 47.80 11.79
phase, the chalcopyrite particles attached onto dispersed bubble 7 93.53 1.0 0.048 4.05 1.90 11.58 2.82 42.19 10.17
surfaces are transported to the froth phase. As a result of the quick 8 94.38 1.8 0.053 5.67 2.10 13.68 2.65 52.50 9.73
drainage of water in the froth, bubbles start to attach to one another. 9 94.40 1.4 0.043 5.48 2.00 13.50 7.61 50.34 10.97
Given that bubbles are densely packed in the froth phase, each la- 10 94.42 1.4 0.043 5.35 2.00 16.99 8.30 53.58 14.08
11 95.44 1.8 0.051 5.12 1.90 13.75 3.04 44.99 9.18
mella is shared by two neighbouring bubbles. The particles original- 12 94.57 1.4 0.043 5.65 1.90 15.49 6.17 50.22 9.64
ly attached on dispersed bubble surfaces redistribute on lamellae. 13 78.51 1.0 0.048 3.93 2.10 10.57 3.83 36.03 8.90
Consequently, the bubble loading in the froth phase is double that 14 92.35 1.8 0.053 5.39 2.30 12.65 10.98 44.19 13.19
in the pulp phase. 15 95.05 1.0 0.048 3.80 1.80 11.82 12.74 33.81 18.86
16 95.48 1.8 0.051 5.30 2.10 9.02 10.95 31.49 14.14
Bubble loading in the pulp phase can be calculated as the ratio of
17 91.87 1.8 0.053 4.40 2.60 4.17 6.91 23.51 2.52
the mass flow rate of chalcopyrite entering the froth phase to the 18 94.52 1.4 0.043 5.98 2.00 16.01 6.50 56.09 12.08
bubble surface area flow rate entering the froth phase. The mass 19 85.35 1.4 0.043 6.15 2.10 16.84 5.30 56.43 9.83
flow rate of chalcopyrite entering the froth phase is equal to the 20 81.38 1.0 0.048 1.99 1.80 1.97 9.57 11.34 3.69
mass flow rate of chalcopyrite in the final flotation concentrate 21 90.72 1.0 0.048 3.01 2.90 4.79 15.97 18.85 9.16
22 85.26 1.0 0.048 2.16 1.30 4.79 21.31 12.21 26.01
when the drainage of chalcopyrite in the froth phase is negligible. 23 91.63 1.4 0.043 2.96 0.80 5.46 12.71 21.15 10.06
This is a reasonable assumption when the final flotation recovery 24 92.11 1.4 0.043 6.22 2.10 17.35 4.37 58.68 13.20
is close to 100%. As shown in Eq. (2), flotation recovery (R) is deter- 25 76.62 0.6 0.058 1.57 1.20 5.19 13.34 14.93 18.11
mined by both pulp recovery (Rc) and froth recovery (Rf) [26]. Froth 26 93.04 1.4 0.043 5.96 2.30 16.66 6.78 56.20 12.38
27 89.05 1.4 0.042 6.02 1.90 15.78 7.71 59.06 11.67
recovery must be close to 100% when flotation recovery is close to
28 94.37 1.4 0.043 5.69 3.10 16.44 12.38 58.06 13.58
100%, and therefore drainage from the froth will be minimal. Once 29 95.89 1.4 0.057 5.72 2.10 12.78 13.17 43.06 15.57
chalcopyrite particles are recovered into the froth phase, they will 30 97.26 2.2 0.048 6.64 2.50 17.37 6.58 63.15 12.66
survive to be in the final concentrate. As shown in Table 2, recover- 31 93.99 1.4 0.043 6.00 2.20 12.19 11.34 48.17 10.93
ies in the flotation tests were high (mostly N 85%). It is therefore a 32 89.71 1.4 0.043 4.74 1.50 7.27 1.74 37.35 3.52
33 92.87 1.4 0.043 5.24 2.90 11.45 6.53 49.11 4.62
reasonable assumption that the mass flow rate of chalcopyrite in
 C. Li et al. / Powder Technology 294 (2016) 55–65 59

there is likely to be chalcopyrite drainage from the froth). However, it is the plateau borders (as silica is not hydrophobic). This can be calculated
believed that this method provides a reasonable relative estimate of the based on the concentrate mass flow rate and the copper grade. The vol-
fraction of lamellae covered by solids between the tests. umetric flow rate of silica in the concentrate (Vs) is determined by
Eq. (10) by considering the silica density (ρs = 2.65 g/cm3). The water
2:89∅c  G volumetric flow rate in the concentrate (Vw) is measured. Finally, the
LB ¼ 2  ð5Þ
100  φp solids concentration in the plateau borders (εs) can be calculated using
Eq. (11).
To simplify the analysis, the shape of the solid particles on the lamel-
lae is assumed to be spherical. The mean particle size of the solids on the ∅c  ð100−2:89GÞ
Vs ¼ ð10Þ
lamellae (P50) is assumed to be the same as that of the sized concen- 100  ρs
trate. It is assumed that the particles exhibit a monolayer distribution
on the lamellae. Hence, the area of a particle occupying the lamella is Vs þ Vc
equal to its projected area. The fraction of the lamella covered by solids εs ¼ ð11Þ
Vs þ Vc þ Vw
(α) can be calculated by Eq. (6). Density in this equation is required to
convert the mass of chalcopyrite loaded on bubbles into a project area
(ρc = 4.18 g/cm3).
4. Results
30; 000LB
α¼ ð6Þ 4.1. Froth rheology
2ρc  P50

The measured froth rheology data were vane speed and torque as
shown in Fig. 5. This data was converted to shear stress and shear strain
3.3. Solids volume fraction in the plateau borders and vertices
to create the standard shear stress shear rate rheograms for each exper-
iment using a method recently proposed for doing these calculations
The solids volume fraction in the plateau borders and vertices is
when using a vane style rheometer surrounded by a tube [21].
determined by the ratio of the volume of solid particles to the total
The changing slope of the torque versus vane speed relationships
volume (including solids and water). It is equal to the ratio of the solids
is an indication that the flotation froths created in the test program
volumetric flow rate (hydrophilic and hydrophobic particles) to the
are exhibiting non-Newtonian shear-thinning behaviour. Thus the appar-
total volumetric flow rate (hydrophilic and hydrophobic particles and
ent viscosity of the froth increases as the shear rate applied decreases. As
water) in the plateau borders and vertices.
the shear rate is expected to be related to froth velocity which is not con-
The volumetric flow rate of hydrophobic particles in the plateau
stant throughout the froth phase, this poses a challenge when one wants
borders and vertices is the overall volumetric flow rate of hydrophobic
to compare the viscosity of different froths and how it changes with froth
particles in the final concentrate minus the flow rate of the particles at-
properties. What viscosity should be compared?
tached on bubble surfaces. The latter is directly determined by the bub-
The Herschel–Bulkley model (Eq. (12)) has been widely used to fit
ble surface area flow rate overflowing the concentrate lip and the
foam rheograms [29–31]:
bubble loading. The bubble loading is obtained from Eq. (5): therefore,
the bubble surface area flow rate overflowing the lip (φf) is needed.
τ ¼ τ y þ μ  γ_
n
The volumetric flow rate of air overflowing the lip (Va) can be deter- ð12Þ
mined by Eq. (7), where εf is the gas hold-up in the froth, vf represents
the surface froth velocity over the lip, h is the froth height above the where τ is the shear stress, τy is the yield stress, μ is the consistency
lip, w is the lip width and k represents the velocity profile of the index, γ_ is the shear rate and n is the flow index (dimensionless). The
overflowing froth. When a linear velocity decrease with depth is as- value of n indicates the deviation of the fluid from Newtonian behaviour:
sumed, the average velocity is half of the measured surface velocity when n N 1, the fluid is shear-thickening (i.e. dilatant); when n b 1, it is
(k = 0.5) [28]. shear-thinning (i.e. plastic or pseudo-plastic). When n = 1, the fluid is
Newtonian, and the viscosity (i.e. μ) is a constant [9].
Va ¼ k  ε f  vf  h  w ð7Þ

The Sauter mean bubble diameter in the froth (D32 ) is obtained from
the surface froth image analysis. The bubble surface area flow rate
overflowing the lip can be calculated using Eq. (8). The factor of “1/2” in
Eq. (8) corresponds to the sharing of each lamella by two neighbouring
bubbles. Hence the mass flow rate of chalcopyrite recovered in concen-
trate attached to lamellae (LB ⋅φf) can be calculated. The volumetric flow
rate of chalcopyrite recovered by entrainment in the plateau borders
and vertices (Vc) can be calculated using Eq. (9).

1 6V a
φf ¼  ð8Þ
2 D32

2:89∅c  G−100  LB  φ f
Vc ¼ ð9Þ
100  ρc

Ventura-Medina and Cilliers [11] reported that the water present in

the froth phase is mainly trapped in the plateau borders and vertices;
the contribution from the lamellae is negligible. Hence, the water flow
rate in the plateau borders and vertices is equal to that in the final
concentrate. The silica flow rate in the concentrate is equal to that in Fig. 5. Froth rheology data measured in the 33 flotation tests.
60 C. Li et al. / Powder Technology 294 (2016) 55–65

Eq. (12) was used to fit the data of shear rates and shear stresses col-
lected from each test. These fitted results show that the yield stress of
the froth in these experiments is very low, which may be a consequence
of the measurements being performed in a flowing froth. Thus yield
stress in Eq. (12) can be made to equal zero without any significant ef-
fect on the degree of fit of the results. It was therefore removed from the
equation. In addition, it is apparent viscosity (η), the ratio between the
shear stress and shear rate of a fluid which is the most commonly
used rheological term to evaluate the rheology of non-Newtonian
flow. By dividing both sides of Eq. (9) by shear rate, an equation can
be developed relating the froth's apparent viscosity and the shear rate
being applied (Eq. (13)).

η ¼ μ  γ_
n−1
ð13Þ

Björn, Monja, Karlsson, Ejlertsson and Svensson [32] has suggested Fig. 7. Flow indices calculated for each test showing their 95% confidence interval.
that the consistency index (μ) in Eq. (13) can be used to compare the
viscosity of fluids when the fluids have similar flow indices (n).
The curves of apparent viscosity versus shear rate for all tests are where ni is the flow index and SEni is the standard error of ni. For the
plotted in Fig. 6 on a log–log scale. As shear rate in the froth is less whole set of n, using the calculated average standard error (SEn =
than 4 s−1 in all the tests [20], Fig. 6 only shows the froth flow curves 0.08) for the n value, Eq. (14) is changed to Eq. (15) as shown below:
for shear rate values less than 4 s−1. Apparent viscosity changes signif-
n1 −n2
icantly with shear rate, confirming that the flotation froths were shear- t ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffi : ð15Þ
thinning fluids. The slopes of all the relationships, however, are very 2SEn 2
similar, indicating that the flow index (n) may be able to be considered
a constant.
The t values were calculated in Excel using the function = TINV
If the flow indexes can be considered constant, the variation of froth
(probability, degrees of freedom) (degrees of freedom = sample
rheology in these tests can be represented by the differences in the con-
size − 2). Statistically, for 95% confidence that the n values are signifi-
sistency indices (μ). In order to validate this hypothesis, the standard
cantly different at the t-value of t95, 3 = 3.18, the critical difference for
error of the flow index for each test determined using Eq. (13) was eval-
flow index should be greater than 0.36. Fig. 7 shows that, in general,
uated by using ‘Solver statistics’. The flow index with its error at 95%
the difference between each pair of flow indices is less than 0.36. There-
confidence interval is plotted in Fig. 7, which shows clearly that the
fore, it may be concluded that there is no statistical difference between
flow indices are very similar. Paired t-tests were performed to test the
the n values. As a result, the flow index can be treated as a constant, and
significance of the observed difference between each pair of n values.
the froth rheology can be evaluated by using only the consistency index.
Eq. (14) was employed to calculate the critical difference for a pair of
The effect of the froth properties estimated above on the froth rheology
n to be statistically different with 95% confidence [33].
in the 33 tests (as determined by the consistency index) is considered in
n1 −n2 Section 5.1.
t ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi ð14Þ
SEn1 2 þ SEn2 2 4.2. Calculated froth properties

A summary of the data measured in the flotation experiments to

perform the calculations described in Eqs. (1) and (3) to (11) is given
in Table 2. A summary of the calculated froth properties for each exper-
iment is shown in Table 3. As mentioned previously, there are seven re-
peat tests among the 33 flotation experiments (i.e. Tests 1, 9, 10, 12, 18,
19, and 26). The variation of the calculated froth properties (CoV, the
ratio of the standard deviation to the mean) determined using these re-
peat tests is shown in Table 4. The standard deviation is generally less
than 10% of the mean value.

5. Discussion

5.1. Influence of froth characteristics on froth rheology

5.1.1. Bubble size

Fig. 8 shows a negative relationship was observed between the con-
sistency index obtained from the froth rheology model and the froth
bubble size (D32). There is scatter in this data, which is expected in a sys-
tem where multiple froth characteristics are having a bearing on the re-
sult. According to the regression analysis there is a 99.98% confidence
that the slope of this relationship is negative and non-zero. Therefore
it can be concluded statistically that bubble size is negatively correlated
with froth viscosity.
From a macroscopic viewpoint, with the air volume fraction being
Fig. 6. Froth rheograms determined for various tests [19]. above 0.95 in the froth in all the experiments (see Table 3), the flotation
 C. Li et al. / Powder Technology 294 (2016) 55–65 61

Table 3
Summary of froth properties.

Test εf D32 (cm) α εs μ (Pa·sn)

1 0.9524 0.91 0.014 0.0835 1.47

2 0.9646 0.89 0.0194 0.0807 2.04
3 0.9586 0.81 0.0095 0.1075 1.80
4 0.9736 0.86 0.0128 0.0968 1.92
5 0.9595 1.00 0.0119 0.1058 1.02
6 0.9662 1.23 0.0027 0.0978 2.02
7 0.9484 0.94 0.0059 0.0913 1.44
8 0.9646 1.09 0.0042 0.0872 1.07
9 0.9564 0.92 0.0111 0.0851 1.39
10 0.9531 0.97 0.0118 0.0983 1.40
11 0.9689 1.15 0.0049 0.1004 1.35
12 0.9564 0.97 0.0117 0.0981 1.24
13 0.9564 0.80 0.0084 0.096 1.73
14 0.9698 0.86 0.0118 0.0871 1.10 Fig. 8. Relationship between froth bubble size and the froth consistency index.
15 0.9598 0.82 0.0147 0.1023 1.77
16 0.9789 0.93 0.0075 0.0872 1.22
17 0.9846 0.76 0.0128 0.0585 1.26
and bubble surface deformation. Froth flow is an irreversible process
18 0.9527 0.84 0.0102 0.0911 1.46
19 0.9524 0.84 0.0107 0.0961 1.48 which involves the rearrangement of bubbles. This leads to the appear-
20 0.9863 0.78 0.0094 0.0556 1.59 ance of local velocity gradients in the fluid confined in the lamellae, as
21 0.9776 0.71 0.0154 0.0737 1.84 the fluid is sheared as shown in Fig. 10a [31]. Bikerman [12] noted
22 0.9844 1.01 0.0072 0.1028 0.64
that the lamella was significantly more viscous than the bulk liquid
23 0.9821 1.07 0.0081 0.0778 0.57
24 0.9499 0.94 0.0068 0.0962 1.02
viscosity. Therefore, the shearing friction in the lamellae significantly
25 0.9698 0.90 0.014 0.1013 1.66 dissipates energy and contributes to the froth viscosity. The length of la-
26 0.9517 0.87 0.0108 0.094 1.27 mella in a given volume of froth is determined by the bubble size. A big-
27 0.9493 0.88 0.0121 0.0847 0.98 ger mean bubble size in the froth indicates a lower total lamella length
28 0.9495 0.82 0.0177 0.0848 1.65
and vice versa. This explains the negative correlation between bubble
29 0.9618 0.93 0.0167 0.0878 1.38
30 0.9643 1.00 0.0074 0.0881 1.43 size and froth viscosity.
31 0.9585 0.86 0.0149 0.0774 1.30 Froth flow also involves bubble surface area change. In static circum-
32 0.9682 1.15 0.0043 0.0672 0.88 stances, a lamella is in a force equilibrium composed of the pressure dif-
33 0.956 0.87 0.0191 0.0756 1.58
ference between neighbouring bubbles and the surface tension. In
sheared foam, the analysis of bubble dynamics has shown that the per-
petual formation and disappearance of foam films between colliding
froth structure can be considered to be similar to a dry foam. In two bubbles leads to variation of the bubble surface area around its mean
phase aqueous foam studies, rheology is governed by the air volume value [39]. The viscous dissipation of energy due to surface area change
fraction and the bubble size [34–37]. Air volume fraction has a positive is related to the surface dilatational modulus which defines bubble sur-
correlation with foam rheology, while bubble size influences foam rhe- face rigidity. The bubble surface dilatational modulus is positively corre-
ology negatively. The dominant factor determining foam rheology, tran- lated with surface tension [40,41]. At liquid–air interfaces, surface
sitions from air volume fraction to bubble size as foam dries. At a low air tension results from the greater attraction of water molecules to one an-
volume fraction (Fig. 9a), isolated bubbles stay in a continuous liquid other than to the molecules in the air. The net effect is an inward force at
phase. When a shearing force is applied on the foam, the resistance to the surface which results in the elastic tendency of the liquid to acquire
deformation is mainly determined by the bulk liquid rheology rather the least surface area possible. Bubble deformation involves surface
than the foam texture [30]. As air volume fraction increases (Fig. 9b), stretch as shown in Fig. 10b: at each local point on the surface, the
the foam regime changes from dispersed bubbles to compressed bub- stretch dissipates energy to offset the resistance resulting from surface
bles separated by the lamellae which form the plateau borders and ver- tension. The small bubbles possess high surface tension and, conse-
tices where the liquid is trapped. When an external shearing force is quently, a higher surface dilatational modulus, which results in a greater
applied on this type of foam, the bubbles are deformed and pass over dissipation of energy during deformation. This is another reason why
one another. The liquid trapped in the plateau borders is also sheared froth viscosity is expected to increase as the size of bubbles decrease.
during the flow. Therefore, the foam is expected to become more vis-
cous and its rheological properties determined by the bulk liquid rheol- 5.1.2. Fraction of lamellae covered by solids
ogy and the bubble size [13]. When foam is very dry with a high air Flotation froth is a three-phase regime involving the presence of
volume fraction (Fig. 9c), the volume of liquid trapped in the plateau solid particles. The solid particles are distributed on the bubble surfaces
borders and vertices is negligible, and bubble deformation and the and likely have an effect on the froth rheology. Fig. 11 confirms this hy-
friction between adhered bubbles define the foam rheology. In such pothesis, showing that the greater the coverage of the lamellae with
dry foam, the most important property affecting foam viscosity is the particles, the greater the consistency index and therefore the viscosity
bubble size [38]. of the froth. Regression analysis indicates that there is a 99.64% confi-
From a microscopic viewpoint, the relationship in Fig. 8 can be ex- dence that the slope of this line is positive and non-zero.
plained in terms of the relative motion between neighbouring bubbles The positive correlation between the fraction of lamellae covered by
solids and froth rheology is likely to be due to two reasons:
Table 4
Statistics of the froth properties calculated using the repeat tests. I. The presence of particles on a lamella likely roughens the bubble
surface as shown in Fig. 12. In the sheared froth, the relative motion
Froth property εf D32 (cm) α εs μ (Pa·sn)
between neighbouring bubbles needs extra energy to overcome the
Mean 0.9536 0.90 0.0115 0.0923 1.39 resistance resulting from the friction between the particles attached
STDEV 0.0020 0.06 0.0012 0.0060 0.10 on the bubble surfaces. The higher the proportion of the lamellae
CoV (%) 0.21 6.12 10.87 6.52 6.99
covered, the greater the friction and the more energy is dissipated
62 C. Li et al. / Powder Technology 294 (2016) 55–65

Fig. 9. Characteristics of a foam as the air fraction increases (a) low air fraction, (b) medium air fraction and (c) high air fraction.

by the relative motion; consequently, the more viscous is the froth. distorts the flow field and can therefore be expected to increase the en-
Moreover, the particles on the lamellae will cluster together due to ergy dissipation during flow, and hence the viscosity. For such cases,
hydrophobic attractive force [42] and Van der Waals' forces between Einstein [49] developed a relationship between suspension viscosity
particles. When relative motion occurs between two neighbouring and the volume fraction of solids as shown in Eq. (16).
bubbles, the particle clusters will be disrupted as the energy over-
comes the bonds between particles. The phenomenon is also expect- ηr ¼ 1 þ 2:5εs ð16Þ
ed to dissipate shear energy.
where ηr is the relative viscosity defined as a ratio of suspension viscos-
ity to the viscosity of suspension medium, and εs is the volume fraction
of solids.
II. The presence of hydrophobic particles on the lamellae can lower When the solids volume fraction increases to a moderate level of
surface tension and enhance froth stability [43,44]. A large number about 10%, the average distance between particles becomes approxi-
of particles on a bubble surface stabilize a film by forming a closely mately equal to their average diameter. The interaction between pairs
packed monolayer. This prevents the interfaces from touching and of particles starts to influence the suspension rheology by dissipating
consequent coalescence between bubbles [45], resulting in small shearing energy. The effect of particle interactions on suspension rheol-
bubble sizes. The probability for the bubbles to coalesce decreases ogy is complex and can be generally expressed as a Taylor expansion in
with an increase in the fraction of lamellae covered by solids, and, powers of the solids volume fraction as shown in Eq. (17) [8], where ci
consequently, the bubble size decreases. As discussed in the previ- are expression coefficients.
ous section, the froth becomes more viscous when bubble size is
decreased. ηr ¼ 1 þ 2:5εs þ c2 εs 2 þ c3 εs 3 þ ⋯ ð17Þ

Above a certain solids concentration, the suspension viscosity in-

5.1.3. Solids volume fraction in plateau borders and vertices creases sharply due to the significant energy dissipation resulting
Within the plateau borders and vertices which form between the from friction between particles. The suspension viscosity tends to infin-
packed bubbles of a flotation froth, water and suspended solids exist. ity when the solids volume fraction is close to the maximum where the
In aqueous suspensions, solids concentration by volume can have a con- suspension stops flowing and behaves like a solid body. In general, the
siderable effect on bulk rheology and potentially the solids within the suspension viscosity increases slightly with increasing solid volume
plateau borders may also be having an effect on the froth rheology. fraction below solids concentration of about 0.50; above this value,
However, there was no trend observed between the froth viscosity (in the suspension viscosity increases much more rapidly with increases
terms of the consistency index) and the estimated solids concentration in solids concentration [8,46,50].
in the plateau borders and vertices (Fig. 13). Regression analysis esti- In this study, the solids concentration in the plateau borders was low
mates that there is a 95% confidence that it is not sufficient to conclude and varied in a narrow range between 5.56% and 10.75% (Fig. 13).
a correlation existing between the consistency index and the solid Therefore, it is believed that the contribution of solids concentration to
fraction. the viscosity of the suspension in the plateau borders and consequently
Now this observation may be a consequence of either the bubbles to the whole froth phase was low. This will result in any effect of the
being so tightly packed that the rheology of the suspension between solids volume fraction in the plateau borders on froth viscosity being
the bubbles plays no role or the percent solids in the water phase is masked by the solids surface coverage and bubble size effects. Therefore
too low to be having an effect. no clear trend was observed in Fig. 13.
In a dilute system in which the solids volume fraction is less than
0.05, the collisions between particles can be ignored [8]. In such cases, 5.2. Modelling froth rheology
the inter-particle interaction contributes negligibly to the suspension
rheology. It is known that the suspension viscosity increases with in- According to the above analysis, bubble size and the fraction of lamel-
creasing solids volume fraction [46–48]. The presence of particles lae covered by solids are the two important froth properties affecting the

Fig. 10. A schematic of the processes determining the lamella dynamics in sheared froth.
 C. Li et al. / Powder Technology 294 (2016) 55–65 63

Fig. 11. Effect of the fraction of lamellae covered by solids on the froth consistency index.
Fig. 13. Effect of solids concentration in the plateau borders and vertices on the
consistency index of the froth.
froth rheology. Based on these findings, the literature was reviewed with
the objective of determining an appropriate model structure that could be
used to predict froth rheology based on the underpinning mechanisms. of 0.5. Fig. 14 shows the comparison between the apparent viscosities
Princen and Kiss [10] was found to have developed a model for foam calculated at various shear rates using the refitted Eq. (13) against
and highly concentrated emulsions that predicts apparent viscosity as a those measured experimentally. The degree of fit is acceptable with
function of the foam or emulsion properties. the average deviation being 4.15% and the regression coefficient (R2)
being 0.92.
τy  
η¼ þ 32:0 ε f −0:73 η0 Ca−0:5 ð18Þ In this study, the flotation froth exhibited no significant yield stress.
γ_ The bulk viscosity of the continuous phase is expected to be close to the
liquid media and will not change as the proportion of solids in the water
where τy is yield stress, εf is gas hold-up, η0 is the viscosity of the in the plateau borders was reasonably low in all tests (Fig. 13). The in-
Newtonian continuous phase and Ca is the capillary number. terfacial tension, which determines the interfacial modulus, is likely to
be associated with the proportion of the lamellae which is covered by
η0 R32 γ_ the hydrophobic particles. Hence, Eq. (20) can be converted to Eq. (21).
Ca ¼ ð19Þ
σ
!−0:5
where R32 is the Sauter mean bubble radius and σ is the interfacial   D32 γ_
η ¼ ε f −0:73 ð21Þ
tension. By substituting Eq. (19) into (18), gives: 2k0 α

 
τ0   R32 γ_ −0:5
η¼ þ 32:0 ε f −0:73 : ð20Þ where k0 is the constant representing those parameters of the system
γ_ η0 σ that should not change between the experiments.
Using the froth property data measured in this study (Table 3), the
This is of the same form as the modified Herschel–Bulkley model apparent viscosity of the flotation froth was fitted to that predicted by
originally fitted to the experimentally produced froth rheograms in Eq. (21) by optimising the value of the constant, k0. k0 was found to
this study but includes parameters related to the properties of the equal 1.22 Pa·s·N·cm− 1. It is therefore that Eq. (21) is converted to
froth system — its yield stress (when it is relevant), gas hold-up, the Eq. (22). Fig. 15 shows a comparison between the experimental appar-
bulk viscosity of the continuous phase which will exist in the plateau ent viscosity and that predicted using the model at a number of the
borders, bubble size and interfacial tension. lower shear rate measurements. The deviation of the prediction aver-
Interestingly the flow index is assumed to equal 0.5, which is similar ages 6.23% and the correlation coefficient (R2) is 0.85. It therefore can
to that observed in the flotation froths produced in the experiments in
this study. A number of other researchers have also shown that the
flow index should equal 0.5 [10,30,51–56]. The average value of the
flow indices of these 33 tests is 0.48 with the standard deviation being
0.07. A t-test was used to judge if the average value of the flow indices
is statistically different from 0.5. The standard t-value t95; 32 is 2.04
(32, degree of freedom). It is calculated that the t-value is 1.25 in this
study, which is less than the t-value of t95; 32. It is therefore statistically
valid to assume that the flow index is equal to 0.5. Hence, the froth
rheograms of these 33 tests were refitted using a constant flow index

Fig. 14. Comparison between the experimental and calculated apparent viscosity when
Fig. 12. Schematic of particles on bubble surfaces. n = 0.5.
64 C. Li et al. / Powder Technology 294 (2016) 55–65

model. Interfacial surface tension which is an input to the model is likely

to be affected by frother type and dosage as well as the hydrophobicity
of the particles attached to the bubble lamellae.
Froth rheology is expected to affect the overall flotation process by
affecting froth residence time. This will be investigated in future work.

Nomenclature
A cell cross-sectional area (cm2)
Ca capillary number
D32 Sauter mean bubble diameter in the pulp (cm)
D32 Sauter mean bubble diameter in the froth (cm)
G copper concentrate grade (%)
h froth height above discharge lip (cm)
Jg superficial gas velocity (cm/s)
k constant that accounts for the velocity profile of the
overflowing froth
Fig. 15. Comparison between the experimental apparent viscosity and that predicted at k0 regression constant (Pa·s·N·cm−1)
various shear rates for all the tests. LB bubble loading in the froth (g/cm2)
n flow index (−)
be concluded that 85% of the deviation in the results can be accounted P50 mean particle diameter of chalcopyrite measured in the con-
for by the model, which is considered good. Most of the parameters centrate (μm)
driving the froth rheology are being taken into consideration. R flotation recovery (%)
R32 Sauter mean bubble radius in the froth (cm)
!−0:5 Rc pulp recovery (%)
  D32 γ_
η ¼ ε f −0:73 ð22Þ Rf froth recovery (%)
2:44α
Sb bubble surface area flux in the pulp (cm2/cm2·s)
vf velocity of the upper surface of the overflowing froth (cm/s)
Va volumetric flow rate of air overflowing the lip (cm3/s)
6. Conclusions VA volumetric flow rate of air entering the cell (cm3/s)
Vc volumetric flow rate of chalcopyrite that is entrained in the
The effects of froth properties on froth rheology were investigated plateau borders and vertices (cm3/s)
under various flotation conditions. The froth properties were calculated Vcs concentrate slurry flow rate (cm3/s)
based on several assumptions. Vs volumetric flow rate of silica overflowing the discharge lip
The froths exhibited shear-thinning behaviour that can be described (cm3/s)
well using a power law model defined by a consistency index and a Vw volumetric flow rate of water overflowing the discharge lip
flow index. No statistical difference was found between the fitted flow (cm3/s)
indices in the different tests. Therefore, the consistency index, which is in- w lip width (cm)
dependent of shear rate, was used to evaluate the froth rheology. α fraction of lamellae covered by solids
The understanding of the rheology of two-phase fluids (i.e. aqueous γ_ shear rate (s−1)
foams or suspensions) was consulted (and expanded upon) with the εf gas hold-up in the froth
objective of understanding the rheology of three-phase flotation froths. εs solids volume fraction in the plateau borders
The results showed that the measured froth rheology could be related to μ consistency index (Pa·sn)
the estimated froth properties in each experiment. Bubble size had a ρc chalcopyrite density (g/cm3)
negative correlation with froth viscosity, which is in line with that ob- ρs silica density (g/cm3)
served in two phase dry foams. In a sheared froth, relative motion be- σ interfacial tension (N/cm)
tween bubbles occurs, which results in bubble deformation. The τ shear stress (Pa)
energy dissipation rate is determined by the friction between bubbles τy yield stress (Pa)
and the bubble surface dilatational modulus, both of which are nega- φf bubble surface area flow rate overflowing the lip (cm2/s)
tively correlated to the bubble size. The fraction of lamellae covered φp bubble surface area flow rate in the pulp (cm2/s)
by solids was found to positively influence the froth viscosity. The pres- ∅c solid flow rate of concentrate by mass (g/s)
ence of solid particles stabilizes froth and consequently generates small η apparent viscosity (Pa·s)
bubble sizes. In addition, the presence of particles on the lamellae will η0 viscosity of the Newtonian continuous phase in a foam or
be expected to increase the surface rigidity and bubble interfacial ten- emulsion (Pa·s)
sion and therefore more energy will be dissipated as the bubbles are ηr relative viscosity
sheared, increasing viscosity proportionally with bubble loading. No
clear relationship between froth viscosity and the solids volume fraction Acknowledgments
in the plateau borders and vertices was observed and this was attribut-
ed to the percent of solids being low and not varying significantly. The authors would like to thank the AMIRA P9P (013908) sponsors
A model structure was developed which enables prediction of the for the project funding. The authors also gratefully acknowledge the
apparent viscosity of a three phase flotation froth based on froth bubble valuable help provided by Mr. Martin Harris from the University of
size, the proportion of the lamellae covered by particles and the shear Cape Town for his advice on froth rheology as well as Professor Tim
rate applied. This model was able to adequately predict the froth rheol- Napier-Munn of the JKMRC who assisted with the statistical analysis
ogy measured in the tests performed in this study. It is hoped that this of the results, Dr. Sam Morar for analysing the froth images produced
model structure can be fitted to any flotation system to enable predic- in this work and Professor J.P. Franzidis for proof reading this paper.
tion of froth rheology. More work, however, is required to validate The assistance of Jon Worth and others in the pilot plant who helped
this hypothesis. There may also be the potential to further refine this set up the test rig is also gratefully acknowledged.
 C. Li et al. / Powder Technology 294 (2016) 55–65 65

Reference [30] B. Herzhaft, S. Kakadjian, M. Moan, Measurement and modeling of the flow behavior of
aqueous foams using a recirculating pipe rheometer, Colloids Surf. A Physicochem. Eng.
[1] X. Zheng, J.P. Franzidis, E. Manlapig, Modelling of froth transportation in industrial Asp. 263 (2005) 153–164.
flotation cells: part I. Development of froth transportation models for attached [31] N.D. Denkov, S. Tcholakova, K. Golemanov, K.P. Ananthpadmanabhan, A. Lips, The
particles, Miner. Eng. 17 (2004) 981–988. role of surfactant type and bubble surface mobility in foam rheology, Soft Matter
[2] M. Zanin, E. Wightman, S.R. Grano, J.P. Franzidis, Quantifying contributions to froth 5 (2009) 3389–3408.
stability in porphyry copper plants, Int. J. Miner. Process. 91 (2009) 19–27. [32] A. Björn, P. Monja, A. Karlsson, J. Ejlertsson, B. Svensson, Rheological characteriza-
[3] N. Barbian, J.J. Cilliers, S.H. Morar, D.J. Bradshaw, Froth imaging, air recovery and tion, Biogas, InTech, 2012.
bubble loading to describe flotation bank performance, Int. J. Miner. Process. 84 [33] T.J. Napier-Munn, Statistical Methods for Mineral Engineers—How to Design Experi-
(2007) 81–88. ments and Analyse Data, 2 ed. Julius Kruttschnitt Mineral Research Centre, Brisbane,
[4] J.G. Wiese, P.J. Harris, D.J. Bradshaw, The effect of increased frother dosage on froth Australia, 2014.
stability at high depressant dosages, Miner. Eng. 23 (2010) 1010–1017. [34] S.J. Cox, A viscous froth model for dry foams in the Surface Evolver, Colloids Surf. A
[5] S. Farrokhpay, The significance of froth stability in mineral flotation—a review, Adv. Physicochem. Eng. Asp. 263 (2005) 81–89.
Colloid Interf. Sci. 166 (2011) 1–7. [35] A.H. Falls, J.J. Musters, J. Ratulowskl, Apparent viscosity of foams in homogeneous
[6] F.N. Shi, X.F. Zheng, The rheology of flotation froths, Int. J. Miner. Process. 69 (2003) bead packs, Soc. Petrol. Eng. J. 4 (1989) 155–164.
115–128. [36] P.C. Harris, Effects of texture on rheology of foam fracturing fluids, Soc. Petrol. Eng. J.
[7] S. Farrokhpay, The importance of rheology in mineral flotation: a review, Miner. 4 (1989) 249–257.
Eng. 36–38 (2012) 272–278. [37] S.A. Khan, C.A. Schnepper, R.C. Armstrong, Foam rheology: III. Measurement of shear
[8] J. Mewis, N.J. Wagner, Colloidal Suspension Rheology, Cambridge University Press, flow properties, J. Rheol. 32 (1988) 69–92.
Cambridge, New York, 2012. [38] A. Bronfort, H. Caps, Faraday instability at foam–water interface, Phys. Rev. E: Stat.
[9] H.A. Barnes, J.F. Hutton, K. Walters, An Introduction to Rheology, Elsevier, Amsterdam, Phys., Plasmas, Fluids, 86 (2012) 1–5.
1989. [39] S. Tcholakova, N.D. Denkov, K. Golemanov, K.P. Ananthapadmanabhan, A. Lips, The-
[10] H.M. Princen, A.D. Kiss, Rheology of foams and highly concentrated emulsions IV. An oretical model of viscous friction inside steadily sheared foams and concentrated
experimental study of the shear viscosity and yield stress of concentrated emul- emulsions, Phys. Rev. E: Stat. Phys., Plasmas, Fluids 78 (2008) 1–18.
sions, J. Colloid Interface Sci. 128 (1989) 176–187. [40] S. Costa, R. Höhler, S. Cohen-Addad, The coupling between foam viscoelasticity and
[11] E. Ventura-Medina, J.J. Cilliers, A model to describe flotation performance based on interfacial rheology, Soft Matter 9 (2013) 1100–1112.
physics of foams and froth image analysis, Int. J. Miner. Process. 67 (2002) 79–99. [41] M. Karbaschi, M. Lotfi, J. Krägel, A. Javadi, D. Bastani, R. Miller, Rheology of interfacial
[12] J.J. Bikerman, Foams, Springer-Verlag, New York, 1973. layers, Curr. Opin. Colloid Interface Sci. 19 (2014) 514–519.
[13] D. Wang, Q. Hou, Y. Luo, Y. Zhu, H. Fan, Blocking ability and flow characteristics of [42] G.V. Franks, H.H. Li, J.-P. O'Shea, G.G. Qiao, Temperature responsive polymers as
nitrogen foam stabilized with clay particles in porous media, J. Dispers. Sci. Technol. multiple function reagents in mineral processing, Adv. Powder Technol. 20 (2009)
36 (2014) 170–176. 273–279.
[14] P.C. Harris, V.G. Reidenbach, High-temperature rheological study of foam fracturing [43] D. Tao, G.H. Luttrell, R.-H. Yoon, A parametric study of froth stability and its effect on
fluids, J. Pet. Technol. 39 (1987) 613–619. column flotation of fine particles, Int. J. Miner. Process. 40 (2000) 25–43.
[15] A.B.J. Kroezen, G.J. Wassink, C.A.C. Schipper, The flow properties of foam, J. Soc. Dye. [44] Z. Aktas, J.J. Cilliers, A.W. Banford, Dynamic froth stability: particle size, airflow rate
Colour. 104 (1988) 393–400. and conditioning time effects, Int. J. Miner. Process. 87 (2008) 65–71.
[16] B. Herzhaft, Rheology of aqueous foams—a literature review of some experimental [45] G. Johansson, R.J. Pugh, The influence of particle size and hydrophobicity on the sta-
works, Oil Gas Sci.Technol. 54 (1999) 587–596. bility of mineralized froths, Int. J. Miner. Process. 34 (1992) 1–21.
[17] G.J. Hirasaki, J.B. Lawson, Mechanisms of foam flow in porous media: apparent [46] P.K. Senapati, B.K. Mishra, A. Parida, Modeling of viscosity for power plant ash slurry
viscosity in smooth capillaries, Soc. Petrol. Eng. J. 25 (1985) 176–190. at higher concentrations: effect of solids volume fraction, particle size and hydrody-
[18] M.A. Bos, T.V. Vliet, Interfacial rheological properties of adsorbed protein layers and namic interactions, Powder Technol. 197 (2010) 1–8.
surfactants: a review, Adv. Colloid Interf. Sci. 91 (2001) 437–471. [47] A.A. Zaman, A.L. Fricke, Correlations for viscosity of kraft black liquors at low solids
[19] M. He, Y. Wang, E. Forssberg, Parameter studies on the rheology of limestone concentrations, AICHE J. 40 (1994) 187–192.
slurries, Int. J. Miner. Process. 78 (2006) 63–77. [48] S. Farrokhpay, G.E. Morris, D. Fornasiero, P. Self, Stabilisation of titania pigment par-
[20] C. Li, S. Farrokhpay, K. Runge, F. Shi, Determining the significance of flotation vari- ticles with anionic polymeric dispersants, Powder Technol. 202 (2010) 143–150.
ables on froth rheology using a central composite rotatable design, Powder Technol. [49] A. Einstein, Investigations on the Theory of the Brownian Movement, Dover, New
287 (2016) 216–225. York, 1956.
[21] C. Li, S. Farrokhpay, F. Shi, K. Runge, A novel approach to measure froth rheology in [50] J.S. Chong, E.B. Christiansen, A.D. Baer, Rheology of concentrated suspensions, J.
flotation, Miner. Eng. 71 (2015) 89–96. Appl. Polym. Sci. 15 (1971) 2007–2021.
[22] S.H. Morar, M. Harris, D. Bradshaw, The use of machine vision to predict flotation [51] S.P.L. Marze, D. Langevin, A. Saint-Jalmes, Aqueous foam slip and shear regimes
performance, Miner. Eng. 36–38 (2012) 31–36. determined by rheometry and multiple light scattering, J. Rheol. 52 (2008)
[23] F. Contreras, J. Yianatos, L. Vinnett, On the froth transport modelling in industrial 1091–1111.
flotation cells, Miner. Eng. 41 (2013) 17–24. [52] S. Tcholakova, I. Lesov, K. Golemanov, N.D. Denkov, S. Judat, R. Engel, T. Danner,
[24] J. Yianatos, F. Contreras, F. Díaz, Gas holdup and RTD measurement in an industrial Efficient emulsification of viscous oils at high drop volume fraction, Langmuir 27
flotation cell, Miner. Eng. 23 (2010) 125–130. (2011) 14783–14796.
[25] M.A. Vera, J.P. Franzidis, E.V. Manlapig, Simultaneous determination of collection [53] R. Foudazi, I. Masalova, A.Y. Malkin, Flow behaviour of highly concentrated emul-
zone rate constant and froth zone recovery in a mechanical flotation environment, sions of supersaturated aqueous solution in oil, Rheol. Acta 50 (2010) 897–907.
Miner. Eng. 12 (1999) 1163–1176. [54] G. Gutiérrez, M. Matos, J.M. Benito, J. Coca, C. Pazos, Preparation of HIPEs with con-
[26] J.A. Finch, G.S. Dobby, Clolumn Flotation, Pergamon Press, London, UK, 1990. trolled droplet size containing lutein, Colloids Surf. A Physicochem. Eng. Asp. 442
[27] B.k. Gorain, J.P. Franzidis, E.V. Manlapig, Studies on impeller type, impeller speed (2014) 111–122.
and air flow rate in an industrial scale flotation cell. Part 4: effect of bubble surface [55] B. Dollet, C. Raufaste, Rheology of aqueous foams, C. R. Phys. 15 (2014) 731–747.
area flux on flotation performance, Miner. Eng. 10 (1997) 367–379. [56] G. Ovarlez, S. Rodts, A. Ragouilliaux, P. Coussot, J. Goyon, A. Colin, Wide-gap Couette
[28] J. Cilliers, R.A. Asplin, E.T. Woodburn, Kinetic flotation modelling using froth imaging flows of dense emulsions: local concentration measurements, and comparison
data, in: J.A. Laskowshi, E.T. Woodburn (Eds.), Frothing in Flotation II, Gordon and between macroscopic and local constitutive law measurements through magnetic
Breach Science Publishers, The Netherlands, 1998. resonance imaging, Phys. Rev. E: Stat. Phys., Plasmas, Fluids 78 (2008) 1–13.
[29] A. Jäsberg, P. Selenius, A. Koponen, Experimental results on the flow rheology of
fiber-laden aqueous foams, Colloids Surf. A Physicochem. Eng. Asp. 473 (2015)
147–155.