Quantifying rheological and fine particle attachment contributions to coarse particle recovery in flotation

Minerals Engineering 39 (2012) 89–98

Contents lists available at SciVerse ScienceDirect

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

Quantifying rheological and fine particle attachment contributions to coarse particle recovery in flotation D. Xu a, I. Ametov a, S.R. Grano b,⇑ a b

Ian Wark Research Institute, The ARC Special Research Centre for Particle and Material Interfaces, University of South Australia, Mawson Lakes, SA 5095, Australia Institute for Mineral and Energy Resources, The University of Adelaide, SA 5005, Australia

a r t i c l e

i n f o

Article history: Received 20 September 2011 Accepted 11 July 2012 Available online 4 October 2012 Keywords: Desliming Fine particle Coarse particle Rheology

a b s t r a c t This study focused on the flotation behaviour of very coarse quartz particles in the presence of fine silica and alumina, both of which were used as pulp viscosity modifiers. A decrease in the contact angle of the coarse quartz particles, caused by the attachment of fine particles was believed to be the principal mechanism accounting for the noted depression. Only small surface coverage of attached fine particles may dramatically decrease the quartz particle recovery because the flotation behaviour of the coarse particles was very sensitive to particle hydrophobicity, e.g. less than 5% surface coverage is able to decrease the contact of particles from 83° to 81° and causes a decrease in recovery from 60% to 20%. The effect of removing the fine particles from the pulp, by the process known as desliming, on the flotation behaviour of coarse quartz particles was also investigated. The results showed that desliming is beneficial for the recovery of coarse quartz particles. Furthermore, the recovery of coarse quartz particles attached with fine particles can be restored by conducting flotation in high viscosity medium where glycerol was used as the viscosity modifier. Ó 2012 Published by Elsevier Ltd.

1. Introduction Detachment of particles from bubbles is one of the key issues responsible for the low recovery of coarse particles. For coarse particles attached to bubbles, the particle–bubble aggregates must withstand the various forces which are operational in the flotation cell to be successfully transported to the pulp/froth interface. A property-based flotation model, developed at the Ian Wark Research Institute (The Wark flotation model) (Duan et al., 2003; Pyke et al., 2003) suggests that a key parameter which directly controls the stability of a particle–bubble aggregate and which quantifies the mean shear forces acting on the bubble–particle aggregate is the mean turbulent energy dissipation. A decrease in the mean turbulent energy dissipation throughout a flotation cell may benefit the recovery of coarse particles due to the reduction of shear forces acting on the particles attached to the bubbles, increasing the stability of the bubble–particles aggregates. Other studies have shown that increasing the viscosity of the pulp results in a decrease in turbulent energy dissipation in a flotation cell (Kitano et al., 1981; O’Connor et al., 1990). It has also been suggested that slurry rheology is an important factor for flotation due to its marked effect on cell hydrodynamics, including gas dispersion throughout the cell (O’Connor et al., 1990; Deglon et al., 2007). O’Connor et al. (1990) found that a decrease in ⇑ Corresponding author. Tel.: +61 0 8 83130626; fax: +61 8 8303 8030. E-mail address: [email protected] (S.R. Grano). 0892-6875/$ - see front matter Ó 2012 Published by Elsevier Ltd. http://dx.doi.org/10.1016/j.mineng.2012.07.003

pulp viscosity in a viscous slurry resulted in a decrease in the bubble size due to increased turbulence in the cell. However, Deglon et al. (2007) found the opposite trend for the change in bubble size, in a study of Bindura nickel ore slurries. According to Deglon et al. (2007), the bubble size decreases with an increase in the solids concentration. Deglon et al. (2007) proposed that the decrease in bubble size was due to the high yield stress in the slurry, which caused a more concentrated energy dissipation near the impeller and leads to the production of small bubbles. A decrease in the gas hold up was also attributed to the high yield stress of the slurry, which prevents the dispersion of bubbles through the cell (Deglon et al., 2007). An increase in slurry viscosity may be achieved by increasing the percent solids, particularly using fine particles as the viscosity modifier. An example of the effect of particle concentration on the rheology of titanium dioxide suspensions was reported by Yang et al. (2001). At relatively low volume fraction of the titanium dioxide (U = 0.109), the suspension shows Newtonian behaviour, i.e. the viscosity is independent of the shear rate. An increase in the solids volume fraction to U = 0.174 results in shear-thinning behaviour, with the viscosity decreasing with an increase in shear rate. With a further increase in the solids volume fraction the rheological behaviour of the suspension remained shear-thinning, but the apparent viscosity values increase considerably by almost three orders of magnitude at U = 0.431 (Yang et al., 2001). Similar trends in rheological behaviour was observed for slurries of dolomite (Deglon et al., 2007), galena (Gao and Forssberg, 1993;

90

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

Wang and Forssberg, 1995), quartz (Prestidge, 1997a,b) and coal (Tangsathitkulchai, 2003), though changes in flow behaviour occur at different volume concentrations of the particles. One of the drawbacks of using fine particles to modify the viscosity is the possible interaction between coarse and fine particles, which may cause a decrease in the flotation recovery of the coarse particles by reducing the hydrophobicity of the coarse particles. Thus, fine particles may modify the flotation behaviour of coarse particles through both viscosity modification and fine particle attachment to coarse particles. This paper investigates the effect of fine particles on the flotation of coarse particles through rheological modification and/or fine particle attachment mechanisms. 2. Experimental 2.1. Material Samples of quartz (GEO Discoveries, Australia) were ground and screened to two coarse size fractions of interest, namely 150– 300 lm and 600–850 lm. Each size fraction of ground quartz were cleaned using the procedure outlined by Pashley and Kitchener (1979). Namely, the particles were washed in concentrated hydrochloric acid three times (2 h each time) and then rinsed with copious amounts of Milli-Q water several times until the pH value of the Milli-Q water (5.6) was restored. The particles were then immersed in a 30% NaOH solution at 60 °C for 1 min, followed by the same rinsing procedure. The particles were then dried in a clean oven at 110 °C overnight and stored in capped bottles in a desiccator under vacuum. Trimethylchlorosilane (TMCS) solutions in cyclohexane were used for particle methylation (Pashley and Kitchener, 1979). Since TMCS readily reacts with water, the methylation reaction was performed in a glove box under nitrogen atmosphere. Different concentrations of TMCS were prepared by diluting the required volumes of TMCS in cyclohexane. The cleaned quartz were weighed into a beaker and heated in an oven at 110 °C overnight to remove the physisorbed moisture. Particles with various contact angles were obtained using solutions of different TMCS concentrations and reaction time. All glassware was cleaned and dried before use. Fine alumina (Hydral 710, Alcoa of Australia Limited) and silica (Sigma–Aldrich Inc., USA) were used to adjust the pulp viscosity. The choice of these particles is based on the fact that at pH < 9, alumina is positively charged (Johnson et al., 2000), and may interact with coarse quartz particles. In contrast, silica has a negative zeta potential in the 3–9 pH range (Ametov and Prestidge, 2004), and is not expected to interact with quartz. The condition of pH 9, i.e., the pHIEP of alumina was used because particle interaction, and consequently viscosity, is most significant at this pH value. The aggregates may facilitate the attachment. Thus, rheological and fine particle attachment contributions to flotation response may be discerned. To contrast the effect of particle interaction in the case of the alumina particles, silica was also examined as a model system that would exhibit lower particle interaction under these conditions. Therefore, using alumina and silica, the effect of colloidal particles on the viscosity of the suspending medium and the flotation behaviour of the coarse quartz may be considered, as a first approximation, as an investigation for the cases of interacting and non-interacting fine particles. Characterisation of fine alumina and silica is presented in the next section. 2.2. Characterisation of fine particles 2.2.1. Zeta potential of alumina and silica particles Zeta potential of silica and alumina was determined from the particle dynamic mobility using the Nano-ZS Zetasizer instrument

(Malvern Instruments Ltd., Worcestershire, UK) in electrophoretic light scattering mode for dilute particle suspensions. Dilute silica and alumina suspensions were prepared at 0.5 wt.% solids, in 103 M KCl and dispersed with a magnetic stirrer for 30 min. The suspension was allowed to stand for 5 min, and the colloidal particles (<5 lm in size) in the supernatant were siphoned off for the zeta potential measurement. The suspension pH was altered to a desired value with HCl and KOH solutions, and allowed to equilibrate for 10 min before the samples were directly injected into a disposable capillary cell for zeta potential measurements. The measurements were performed over the pH range 5–9.5 for alumina. The zeta potential of the silica was investigated over pH range 3–9. 2.2.2. Particle size distribution of alumina and silica The particle size distribution of alumina and silica was determined by laser diffraction using a Mastersizer 2000 (Malvern Instruments Ltd., UK). The basic particle size sensor comprises an optical measurement unit which supplies information to a computer to process data and perform the analysis. Ultrasonication for 5 min and a polyphosphate dispersant (i.e., Calgon) were used to achieve full dispersion of both alumina and silica particles. 2.2.3. Slime coating of fine particle on quartz surface Scanning electron microscopy (a PHILIPS XL-20 electron microscope) was used to determine the adsorption of fine silica and alumina particles on the surface of coarse quartz particles. The samples were mounted onto the sample holder using double-sided sticky tape and were coated with a thin carbon layer using a vacuum evaporator. 2.2.4. Slurry rheology The rheological behaviour of the silica and alumina suspensions was investigated using a Haake RotoVisco RV1 rheometer (Thermo Electron GmbH, Germany) fitted with the concentric cylinder (Couette) sensor. Each slurry sample (100 cm3) was prepared by adding known weights of fine silica or alumina particles to known volumes of the 103 M KCl solution at pH 9. Prior to the rheology measurement, the suspensions were stirred for 1 h using an overhead stirrer. In the measurement, the shear rate increased from 0 s1 to 1000 s1 (an upward curve) and then decreased down to 0 s1 (a downward curve) in 200 s. The rheological parameters were automatically recorded by a computer. The percentage of fine alumina and silica is shown as volume% in all cases using 2.65 g/cm3 and 3.95 g/cm3 density for silica and alumina respectively to convert the mass of fine particles to corresponding volume. 2.3. Methodology 2.3.1. Flotation Flotation tests on coarse quartz particles were carried out according to the flowsheet shown in Fig. 1. Quartz particles (60 g) of various mean contact angle values and particle sizes were floated in a 1.5 dm3 bottom driven flotation cell. All flotation tests were carried out at an impeller speed of 600 rpm. Air was introduced into the flotation cell at a flow rate of 3.5 dm3/min (Jg = 0.4 cm/s). Dowfroth 250 (170 g/t, 8 ppm in solution) was used as frother. Froth depth was 3 ± 1 cm. Quartz particles were hydrophobised to the target value of contact angle before the flotation test. Four concentrates were collected, cumulatively at 0.5 min, 2 min, 4 min and 8 min. Make-up water was used to the cell to keep the interface at the same level during the flotation tests. The viscosity of the suspending medium was increased by: (i) using glycerol (95% purity)/water mixtures instead of water and

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

Fine Size Gangue Particles

2.3.3. Desliming The flotation recovery of coarse quartz particles generally decreased in the presence of the colloidal silica and alumina particles, presumably due to the adsorption of fine particles onto the surface of the coarse quartz. To investigate whether the adsorption of fines was reversible, a series of desliming tests was carried out. The desliming procedure is schematically presented in Fig. 2. Slurry, containing coarse quartz and fine silica or alumina particles was placed in the flotation cell as per the standard flotation procedure. The coarse quartz particles were allowed to settle. The supernatant, containing fine particles was then decanted, the cell was refilled with the 103 M KCl solution and the flotation test was performed. In a different test, the desliming procedure was repeated twice to increase the amount of removed fine particles. The flotation test was also conducted following the procedure described above.

Glycerol

Flotation

Concentrates

Tailing

Frother 0-0.5 min

2-4 min

0.5-2 min

4-8 min

Fig. 1. The schematic diagram of conventional flotation tests conducted on coarse quartz particles.

(ii) by the addition of the colloidal alumina or silica particles at different volume concentrations. Flotation tests in water and the glycerol/water mixtures were conducted at pH 7. In the case where fine particles were used to increase the viscosity of the suspending medium, 103 M KCl solution was used as a background electrolyte, and the pH of the slurry was maintained at 9. To determine the recovery of coarse quartz particles in the flotation tests in the presence of alumina and silica, the concentrates and tailings were wet screened at 38 lm to remove the fine particles. To estimate the experimental errors, all tests were performed in triplicate.

2.3.2. Experimental flotation rate constant Assuming that froth flotation is a first-order kinetic process, the flotation recovery, Rt, at time t may be described by the following expression:

Rt ¼ Rmax ð1  expðktÞÞ

91

ð1Þ

where k is the flotation rate constant and Rmax is the flotation recovery at an infinite time. A nonlinear least square regression was used to calculate k and Rmax from the best fit of the curve of experimental flotation recoveries versus time using Eq. (1). These flotation rate constants are referred to as the experimental flotation rate constants in the text. Small variations in the flotation recovery versus time data may result in different values of flotation rate constant. To minimise errors associated with the calculations of the rate constant, the tests were carried out in triplicate and the average values of flotation recovery at each time were used to calculate the rate constant.

2.3.4. Contact angle measurement Advancing mean contact angles of the quartz particles in water were measured using DCAT 11/DCAT 11EC, DataPhysics Instruments (Germany). This instrument allows measurements of the contact angle and surface energy of cylindrical samples in accordance with the Washburn technique (Muganda et al., 2011). Eq. (2) was used to calculate the contact angle for two-liquid systems.

cos h1 ¼

c2 l1 q22 ðx2 =tÞ1 cos h2 c1 l2 q21 ðx2 =tÞ2

ð2Þ

where c is the surface tension of the liquid, l is the viscosity of the liquid, and q is the density of the fluids. The subscripts 1, 2 refer to the liquid where the contact angle is determined and calibrating liquid respectively. The contact angle in water was measured using cyclohexane as the calibrating liquid. Thus, in Eq. (2) cos h2 = 1. Given other parameters (c, l, q, and x2/t which could be determined during the measurement), the contact angle may be determined. 3. Results 3.1. Characterisation of the fine alumina and silica 3.1.1. Zeta potential of alumina and silica particles Fig. 3 shows the zeta potential of alumina and silica particles as a function of pH at ionic strength of 103 M KCl. It is apparent that at pH 9, alumina is slightly positively charged (+2 mV) and silica is negatively charged (40 mV). The results are consistent with those reported elsewhere (Ametov and Prestidge, 2004; Johnson et al., 2000). 3.1.2. Particle size distribution of the fine alumina and silica Fig. 4 shows the particle size distribution data for the fine particles of both alumina and silica. Evidently, a broader size distribution for the silica than the alumina particles is observed. The mass mean size (D [4,3]) of the fine silica and alumina particles are 3.09 lm and 3.10 lm respectively. The size distribution shown to

Fig. 2. Experimental procedure used for desliming of coarse quartz particles.

92

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

280

(a)

210

Alumina Silica

25

6% v/v 10% v/v 14% v/v

140 0

-25

-50 2

3

4

5

6

7

8

9

10

Shear Stress (Pa)

Zeta Potential (mV)

50

70

0

(b)

100

9% v/v 20% v/v 36% v/v

pH 75 Fig. 3. Zeta potential of alumina and silica as a function of pH at ionic strength of 103 M KCl.

50 10

25 silica alumina

8

0 0

200

400

600

800

1000

Volume (%)

Shear Rate (1/s) 6 Fig. 5. Shear stress as a function of shear rate and solid volume percentage for (a) alumina and (b) silica suspensions at pH 9 in 103 M KCl solution.

4 0.6

0 0.1

1

10

Particle Size (µm) Fig. 4. Particle size distributions of alumina and silica particles.

characterise the fine particles may not be the actual particle size distribution in the flotation cell due to aggregation or dispersion effects under shear conditions.

Apparent Viscosity (Pa.s)

2 0.5

0.4

Alumina (mean size 3 µm) Silica (mean size 3 µm)

0.3

0.2

0.1

0.0

3.1.3. Rheology of the fine alumina and silica suspensions The flow curves for alumina and silica suspensions as a function of solids percent (v/v) are presented in Fig. 5a and b respectively. The measurements were performed at pH 9 in 103 M KCl solution, at which pH the zeta potential of alumina and silica particles is +2 mV and 40 mV, respectively (Fig. 3). Alumina suspensions exhibit shear-thinning behaviour, i.e. the viscosity of the suspensions decreases with an increase in the shear rate. At the low value of zeta potential, the van der Waals attractive forces are dominant, and alumina particles readily aggregate when the suspension is at rest. Under shear, the large particle aggregates are separated into smaller units and, at very high shear, into individual particles. Such a process rheologically manifests itself as shear-thinning behaviour. Silica suspensions show Newtonian behaviour, i.e., the viscosity of the suspension is independent of the shear rate. At pH 9 in 103 M KCl solution, the zeta potential of the silica particles is 40 mV, and the electrostatic repulsion dominates the inter-particle

0

10

20

30

40

50

Percent Solids (vol. %) Fig. 6. Apparent viscosity (at 500 s1) of fine alumina and silica suspensions as a function solids volume percent in 103 M KCl solution at pH 9. The horizontal line shows the viscosity of interest (30 mPa s) and the corresponding solids percent (v/ v) of silica and alumina.

interactions. The strong electrostatic repulsion ensures good dispersion of particles resulting in Newtonian behaviour. The apparent viscosities (at a shear rate of 500 s1) of the silica and alumina suspensions as a function of solids volume fraction are shown in Fig. 6. Evidently, the apparent viscosity increases with an increase in the solids concentration. In the case of alumina, the increase in the apparent viscosity is much steeper than for silica, also due to the nature of the inter-particle interactions in alumina (the van der Waals attraction) and silica (the electrostatic repulsion) suspension. The aim of this investigation was to use two

93

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

30

(a)

100

(a)

75

2% v/v Alumina 4% v/v Alumina Water only

20

Water only 2% v/v Alumina 3% v/v Silica

25 0

(b)

100 0 3% v/v Silica 30% v/v Silica Water only

20

Quartz Recovery (%)

Quartz Recovery (%)

50 10

10

(b)

80 60 40 20

0 0

3

6

9

20

(c)

Flotation Time (min) Fig. 7. Recovery of coarse quartz particles (600–850 lm) as a function of flotation time and the volume concentration of fine alumina (a) and silica (b) particles. Coarse quartz particles with mean contact angle of 81° in the 600–850 lm size range were used.

types of colloidal particles to increase the pulp viscosity, and then to ascertain the effect of the higher viscosity on the flotation recovery of coarse quartz. Of course, the value of the viscosity attained by the addition of these types of particles is required to be comparable. Two values of the viscosity were selected for investigation, i.e., 12 mPa s and 30 mPa s. From Fig. 6, and 2% (v/v) alumina and 3% (v/v) silica suspensions provide a viscosity of 12 mPa s. To obtain a slurry viscosity of 30 mPa s, 4% (v/v) alumina and 30% (v/v) silica are required, respectively.

15

10

5

0 0

2

4

6

8

Flotation Time (min) Fig. 8. Quartz recovery as a function of flotation time in the presence of fine alumina and silica. Quartz particles in these tests were in the 150–300 lm size range, with the contact angle of (a) 75°, (b) 50° and (c) 18°.

3.2. Effect of fine silica and alumina on coarse quartz flotation The recovery of coarse quartz particles (600–850 lm, mean contact angle 81°) in the presence of fine alumina and silica as a function of flotation time is presented in Fig. 7. The recovery of coarse quartz particles decreased dramatically, from 27% to 2%, for tests in which the slurry contained 2% (v/v) alumina. A further increase in the volume fraction of alumina, to 4%, did not affect the outcome of the flotation tests. The effect of silica at 3% (v/v) was less severe, i.e. the recovery of coarse quartz decreased to 13%. Increasing the concentration of fine silica to 30% (v/v) resulted in a further decrease in coarse quartz recovery to 3%. Evidently, both types of fine particles negatively affected the recovery of coarse quartz particles. In the case of alumina, which is positively charged under the test conditions (pH 9, 103 M KCl), some degree of interaction with the coarse quartz particles was expected. Adsorption of hydrophilic alumina particles onto the surface of hydrophobic quartz may lead to a decrease in the mean contact angle of the coarse quartz surface, and therefore decrease coarse particle recovery. However, a decrease in the flotation recovery of coarse quartz particles was observed in the presence of fine silica, though to a lesser degree. This was not expected due to the surmised electrostatic repulsive force. The effect of the fine silica and alumina on the flotation behaviour of the coarse quartz particles was also investigated as a function of coarse quartz particle size range and mean contact angle. The results of flotation tests for quartz particles in the

150–300 lm and 600–850 lm size ranges are shown in Figs. 8 and 9, respectively. It is worth mentioning that the reason for using two different hydrophobicity ranges for the two size fractions is that for coarser particles, greater hydrophobicity is required to recover the particles. The contact angle ranges shown in the study is the range where a reasonable recovery was obtained for the coarse particles in water only. Thus we can investigate the effect of fine particles on the recovery further. For coarse particles with a reasonably high mean contact angle (75° and 90° for quartz particles in the 150–300 lm and 600– 850 lm size ranges, respectively), the recovery of coarse quartz was not affected by the presence of fine silica and alumina in the pulp. The magnitude of the coarse quartz recovery remained as high as the value obtained in water. In contrast, for quartz particles with lower mean contact angle (50° and 18° in the 150–300 lm size range and 83° for the 600–850 lm size range) the flotation recovery of coarse quartz decreased in the presence of fine alumina and silica. Notably, the effect of the fine alumina on the recovery of quartz was much stronger compared to that of the fine silica although the viscosity values in both cases were the same at 12 mPa s. It should be pointed out that in the case of coarse quartz particles with a contact angle of only 18° (150–300 lm size range) (Fig. 3.6c), the effect of alumina and silica was comparable, i.e., the recovery of quartz decreased to a very low value of 3%.

94

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

100

(a)

100

(a)

water only 3% v/v silica 2% alumina

80

Quartz Recovery (%)

80

60

40 Water only

60

40

Quartz Recovery (%)

2 vol.% Alumina 20

3 vol.% Silica

20

0 100

10

(b)

20

30

40

50

60

70

80

Contact Angle (Degree)

80 100

(b)

water only 2% v/v alumina 3% v/v silica

60

Quartz Recovery (%)

80

40

20

0

60

40

20 0

2

4

6

8

Flotation Time (min) Fig. 9. Quartz recovery as a function of flotation time in the presence of fine alumina and silica. Quartz particles in these tests were in the 600–850 lm size range, with mean contact angle values of (a) 90° and (b) 83°.

In previous work (Xu et al., 2011), it was demonstrated that the recovery of coarse particles increased with an increased medium viscosity, controlled by adding certain amount of glycerol. Particularly, when the medium viscosity is 7.6 mPa s, which is similar to that investigated in this work, 12 mPa s, the recovery of quartz particles with a contact angle of 80° increased from 20% to 70%. It was concluded that an increase in medium viscosity may benefit the flotation recovery of coarse particles. However, in the presence of fine particles, the flotation recovery of coarse particles decreased even though the viscosity was similar. A possible mechanism is discussed further below. The recovery of coarse quartz particles after 8 min of flotation as a function of mean contact angle in the absence and presence of fine silica and alumina is shown in Fig. 10. Evidently, the recovery of the coarse quartz particles in the 150–300 lm size range is less sensitive to the presence of fine alumina and silica (Fig. 10a). The flotation recovery of the quartz particles in the 150–300 lm size range with a mean contact angle of 75° in the presence of fine alumina or silica was as high as in the absence of fine particles, at 90%. For particles with a contact angle of 50° the recovery decreased to 22% in the presence of the fine alumina, and to 68% in the case of fine silica. Particles with low hydrophobicity (h = 18°) exhibited very low recovery (3%) in the presence of either fine silica or alumina. Furthermore, the flotation response of very coarse quartz particles (600–850 lm) to the presence of silica or alumina was

0 68

72

76

80

84

88

92

Contact Angle (Degree) Fig. 10. Effect of fine particles on coarse quartz recovery, coarse particle size. (a) 150–300 lm, (b) 600–850 lm.

noticeably different. In the pulp containing silica particles (3%), the recovery of quartz with mean contact angle of 90° remained virtually the same as in water only (92%). The flotation recovery of quartz particles with a mean contact angle of 83° was 60% in water, and decreased to 50% in the presence of silica (Fig. 10b). There was only 30% recovery at a lower mean contact angle of 81° even in water only. With fine silica added to the pulp the quartz (81o) recovery decreased to 15%. In the presence of fine alumina, the flotation recovery of coarse quartz with a mean contact angle of 90° was as high as in water only or as in the presence of silica (92% in both cases). In contrast, the recovery of coarse quartz particles with mean contact angle of 83° decreased from 60% in water to 10% in the presence of fine alumina (4%). Particles with a mean contact angle of 81° became essentially unrecoverable in the presence of fine alumina particles. It is also shown in Fig. 10 (water only) that the floatability of very coarse particles is very sensitive to the hydrophobicity of the particles. A small decrease in the contact angle (from 83° to 80°) results in a dramatic decrease in the flotation recovery (from 60% to 20%). Increasing the medium viscosity by fine particle addition failed to increase the quartz particle recovery may be due to the adsorption of fine particles on the quartz surface, which resulted in a decrease in particle contact angle. By assuming the

95

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

contact angle of fine hydrophilic particles adsorbed onto the quartz surface is zero, only 4% surface coverage is sufficient to decrease the contact angle of the quartz particles from 83° to 80° according to Cassie’s equation (Cassie and Baxter, 1944). This may explain why the flotation recovery of coarse particles decreased in the presence of the fine particles, which in fact increase the pulp viscosity. The attachment of fine particles may dominate the effect of pulp viscosity for the flotation behaviour of coarse particles.

In the case of silica (3% v/v) suspension (Fig. 11b), the recovery of quartz particles (80o) also decreased, from 20% to 13%. A single stage of desliming did not improve the coarse quartz recovery. However, two stages of desliming process successfully restored the recovery of coarse quartz particles to the original value obtained in water only.

3.3. Effect of removal of fine particles (desliming) on coarse quartz particle recovery

In the previous sections, it was shown that the recovery of coarse quartz decreases in the presence of fine particles, particularly in the case of alumina. It was also proposed that the loss of recovery of coarse particles was due to the decrease in particle hydrophobicity which resulted from the adsorption of fine hydrophilic particles onto the quartz surface. Additionally in previous work (Xu et al., 2011), it was demonstrated that the recovery of coarse quartz particles could be increased by conducting flotation in a high viscosity medium. Moreover, in high viscosity medium coarse quartz particles with lower critical contact angle (Crawford and Ralston, 1988) were able to be recovered (Xu et al., 2011). Therefore, the question now becomes – Is it possible to restore the flotation recovery of coarse quartz particles decreased in the presence of fine alumina and silica by using a high viscosity medium? – is logical and needs to be answered. In this investigation, the coarse quartz particles were conditioned in the presence the fine alumina (2% v/v) or silica (3% v/v) following the flotation procedure described in the previous section. A single stage of desliming was carried out using a flotation cell. In the ‘‘control’’ test, 103 M KCl solution was used to increase the pulp volume in the flotation cell to the required volume and the flotation was conducted. In the test investigating the effect of a high viscosity medium, the calculated amount of glycerol was added to ensure a 50% glycerol/water mixture with the viscosity of 7.6 mPa s, and the flotation test performed.

Since the presence of fine particles had a negative effect on quartz recovery, the effect of their removal in desliming was also investigated. The desliming procedure was outlined in the experimental section above. The effect of desliming on the coarse quartz particles recovery on the size range of 600–850 lm is shown in Fig. 11. In the presence of alumina (2% v/v) (Fig. 11a) the flotation recovery of quartz particles decreased from 60% to 10%. After a single stage of desliming, the recovery of coarse quartz particles increased from 10% to 50%. An additional desliming stage did not result in an appreciable increase in coarse quartz recovery. It seems reasonable to assume that, in the presence of alumina particles in the pulp, the flotation recovery of coarse quartz may always be below that of the original value achieved in the complete absence of alumina particles.

80

(a) Contact Angle 83o

60

3.4. Recovery of coarse quartz in glycerol/water mixture

80 water with 2% alumina single stage de-sliming two stage de-sliming

20

(a) Alumina, Contact Angle 83o

60

40

Quartz Recovery (%)

Quartz Recovery (%)

40

0

(b) Contact Angle 80o

20

water only Single stage de-slimed, in water Single stage de-slimed, in 50% glycerol

20

0

(b) Silica, Contact Angle 80o water only Single stage de-slimed, in water Single stage de-slimed, in 50% glycerol

60

40 10

water with 2% alumina single stage de sliming two stage de sliming

20

0 0

0 0

2

4

6

8

2

4

6

8

Flotation Time (min)

Flotation Time (min) Fig. 11. Effect of desliming on the recovery of coarse quartz particles (600–850 lm) from alumina (a) and silica (b) containing slurries.

Fig. 12. Quartz particle recovery restored by increasing pulp viscosity using 50% glycerol/water mixture for quartz particles in the presence of fine alumina (a) and fine silica (b).

96

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

Table 1 Summary of flotation rate constants and recoveries (Rmax) and water recoveries (Rw) for coarse quartz particles in the 600–850 lm size fraction in the desliming tests. Coarse quartz flotation conditions

Water (No fines) Slime particles Single stage de-slimed, floated in water Single stage de-slimed, floated in 50% glycerol/water mixture

Slime alumina, 2% v/v ° Quartz contact angle 83

Silica, 3% v/v Quartz contact angle 80°

k

Rmax

Rw

k

Rmax

Rw

1.8 0.69 1.27 1.08

59 12 49 65

16 28 23 45

1.5 1.2 0.87 0.79

18 14 14 34

14 18 20 30

Fig. 13. SEM images of quartz particles in the flotation tailings (the contact angle of coarse quartz was 90°). Flotation tests were conducted in the presence of alumina (a) and silica (b) at 2% v/v and 3% respectively. Arrows point to adsorbed particles coarse quartz size range (600–850 lm).

The recovery of quartz particles as a function of flotation time in the ‘‘control’’ test (water) and in the test conducted in the 50% glycerol/water mixture are presented in Fig. 12. The flotation data was fitted using the first order kinetic equation (Eq. (1)). The flotation rate constant and the maximum recovery at infinite time Rmax are summarised in Table 1. It is apparent that the presence of fine particles in the pulp resulted in a decrease in the recovery of coarse quartz, an increase in the water recovery, as well as the flotation rate constant (Table 1). The effect of the fine alumina on flotation of quartz was more pronounced than that of the fine silica. A single stage of desliming, followed by flotation in water, partially restored the recovery of the coarse quartz particles. The flotation rate constant in both cases (for the presence of fine alumina and silica) also increased. It is notable that there is not much difference in water recovery with desliming which indicates that the increase in recovery with desliming is not due to increased entrainment. However, for flotation tests after the single stage desliming but in the presence of the 50% glycerol/water mixture showed that the coarse quartz recovery increased to a value higher than in the ‘‘control’’ tests. In fact, it was higher than the recovery of the coarse quartz in water in the absence of fine particles, i.e. marginally higher for fine alumina particles, and considerably higher in the case of the fine silica (Table 1). In contrast, the flotation rate constant in the tests conducted in the 50% glycerol/water mixture decreased, similarly to what was observed and discussed in previous work (Xu et al., 2011). Furthermore, highest water recovery was observed in the 50% glycerol/water mixture for both cases (alumina and silica) due to the increased pulp viscosity.

4. Discussion 4.1. Effect of slime coating on quartz flotation The flotation results indicate that fine particles in the pulp have a negative effect on the flotation behaviour of the coarse quartz. The fine particles were used to increase the pulp viscosity and, possibly, the flotation recovery of coarse quartz particles. However,

the flotation recovery of quartz decreased. The supposition that the coarse particle recovery may increase in the presence of fine particles is based on previous work that increased viscosity (through addition of glycerol) increased coarse particle recovery (Xu et al., 2011). That the fine particles actually caused depression, while also increasing pulp viscosity, demonstrated that an additional mechanism was at play. It was demonstrated in previous (Fig. 10 water only) and also in previous work (Xu et al., 2011) that the flotation response of coarse particles is very sensitive to the particle hydrophobicity. Calculations discussed previously, suggest that only very low surface coverages of hydrophilic fine particles (of the order of 5%) are required to reduce the contact angle below the critical value depressing flotation. The decrease in the recovery of very coarse particles, even though in higher pulp viscosity, may be due to the fact that hydrophilic fine particles adsorbed on the quartz surface, and consequently reduce the contact angle of the coarse quartz particles to below the critical contact angle. More significant decreases of recovery in the case of fine alumina than that of silica may be expected due to the different nature of surface charge and propensity to adsorb onto the coarse quartz surface. To elucidate the possible mechanism of the decrease in recovery of quartz particles in the presence of fine alumina and silica, scanning electron microscopy (SEM) was used. Coarse quartz particles, collected from the tailings were gently washed with water and studied by SEM. The SEM images of the quartz particles are presented in Fig. 13. Evidently, both types of particles, i.e. alumina (Fig. 13a) and silica (Fig. 13b), adsorb onto the coarse quartz surface. Possibly, a greater amount of alumina is present on the quartz surface compared to silica. Moreover, the majority of the alumina particles adsorb on edges or steps on the quartz surface. This may be due to the higher surface energy of the edges, which are rougher compared to the crystal planes, and possibly provide more adsorption sites. Another possible reason for greater quantities of alumina particles on the quartz surface is the self-aggregation of alumina particles. At pH 9, which is near the iso-electric point for alumina, the surface charge of alumina is low (zeta potential is +2 mV), the attractive

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

van der Waals forces are dominant, and fine particles aggregate. Decreasing the pH to values further away from the iso-electric point may reduce the self-aggregation of alumina, but it may also promote adsorption of alumina onto the coarse quartz surface due to electrostatic interaction. It was reported earlier that both silica and TMCS treated quartz are highly negatively charged (the zeta potential is 40 mV) and, due to the dominant electrostatic repulsion, may not interact under the conditions of the experiment. However, the SEM image in Fig. 13b showed that fine silica particles also adsorb onto the surface of the coarse quartz, although to a lesser extent compared to alumina. Similarly to the fine alumina, fine silica particles also adsorb onto the edges and steps of the quartz surface. Although, the mechanism of adsorption of fine silica on the surface of hydrophobic quartz is not understood, it is clear that only a small surface coverage of either fine alumina or fine silica are required to decrease the contact angle of the coarse quartz particles below the critical value necessary for stable bubble–particle attachment. As previously outlined, the critical contact angle for 600–850 lm size range is of the order of 85°, with decreases in contact angle below 80° able to effectively depress flotation completely. Surface coverage of hydrophilic fine particles less than 9% is able to decrease the contact angle by 5° making the flotation recovery of coarse particles very sensitive to the surface coverage of fine hydrophilic particles. This gives rise to the often noted ‘knife-edge’ behaviour of coarse particles in plant practice. 4.2. Effect of desliming on quartz recovery As shown earlier, the removal of fine particles from the pulp is, to some extent, beneficial for the flotation of coarse particles. A single stage of desliming increased the recovery of coarse quartz particles in the presence of fine alumina, although in the presence of fine silica, it was rather unsuccessful. In contrast, adding a second stage of desliming for the alumina containing pulp was ineffective, but appreciably improved the recovery of quartz in the pulp containing fine silica. The positive effect of desliming may be attributed to detachment of fine particles from the surface of the coarse particles, thus restoring the value of the contact angle. However, it seems that the success of the desliming process depends on the type of the interaction between the fine and coarse particles, as well as on the concentration of fine particles in the pulp and, possibly, turbulence and fluid velocity in the flotation cell (Edwards et al., 1980; Bandini et al., 2001). 4.3. Restoring the recovery of quartz particles using high viscosity medium (50% glycerol/water mixture) In previous work (Xu et al., 2011) it was demonstrated that much higher flotation recovery of coarse particles may be achieved in the high viscosity medium. This effect was attributed to changes in the hydrodynamic parameters in a flotation cell, such as an increase in the mean bubble size, broadening of the bubble size distribution and a decrease in local turbulent energy dissipation and fluid velocity. In addition, it was shown that the critical contact angle, the minimum contact angle required for particles to float, is lower in high viscosity medium. An increase in the medium viscosity reduces the kinetics of the bubble–particle collection process, but at the same time, increases the maximum recovery of coarse particles. The decrease in flotation rate constant in the high viscosity medium may be attributed to the decrease in turbulent energy dissipation and fluid velocity, and consequently the decrease in bubble particle collision frequency. Conversely, low energy dissipation and fluid velocity are favourable for greater stability of bubble–particle aggregates, and therefore the maximum recovery of the coarse particles increases.

97

Adsorption of fines onto the surface of the coarse particles decreases the coarse particle contact angle. In the cases, where the coarse particles become unrecoverable as a result of the adsorbed fine particles, their contact angle decreases to a value lower than the critical contact angle determined in water. Conducting the flotation tests at higher viscosity is beneficial to the recovery of the coarse particles in two ways: (i) the high viscosity medium decreases turbulent energy dissipation in the flotation cell, resulting in an increase in the stability of bubble–particle aggregates and (ii) the contact angle of coarse particles does not need to be as high for the particles to be recovered. In practice, desliming is sometimes used to remove fine particles from the pulp and to reduce the amount of fine hydrophilic particles attached to the coarse particles. This is recognition of the controlling behaviour of the fine particles on coarse particle recovery, and the sensitivity of coarse particles to changes in contact angle. Moreover, desliming may also be accompanied by deliberate changes in pulp viscosity by the introduction of high viscosity media. This combination of approaches has been used to successfully increase the recovery of coarse composite particles from an ore (Farrokhpay et al., 2011). 5. Conclusions The depression of coarse particle recovery was due the attachment of fine hydrophilic particles, which resulted in a decrease in particle mean contact angle. Increasing the pulp viscosity using fine particles failed to benefit the coarse particle recovery due to the dominant effect of fine particle attachment, which was attributed to the fact that the flotation recovery of coarse particle is very sensitive to particle hydrophobicity. Desliming was successful in restoring recovery to a degree probably due to detachment of fine hydrophilic particles and an increase in the contact angle of coarse particles. Further increases in the recovery of coarse particles were apparent with the high viscosity medium using glycerol as a viscosity modifier. Acknowledgement Financial support from AMIRA International, the Australian Research Council and University of South Australia, is gratefully acknowledged. References Ametov, I., Prestidge, C.A., 2004. Hydrophobic interactions in concentrated colloidal suspensions: a rheological investigation. Journal of Physical Chemistry B 108, 12116–12122. Bandini, P., Prestidge, C.A., Ralston, J., 2001. Colloidal iron oxide slime coatings and galena particle flotation. Minerals Engineering 14, 487–497. Cassie, A.B.D., Baxter, S., 1944. Wettability of porous surfaces. Transactions of the Faraday Society 40, 546–551. Crawford, R., Ralston, J., 1988. The influence of particle size and contact angle in mineral flotation. International Journal of Mineral Processing 23, 1–24. Deglon, D.A., Shabalala, N.Z.P., Harris, M.C., 2007. Rheological effects on gas dispersion. Minerals Engineering International Conferences: Flotation 07 Cape Town, Xstrata Process. Duan, J., Fornasiero, D., Ralston, J., 2003. Calculation of the flotation rate constant of chalcopyrite particles in an ore. International Journal of Mineral Processing 72, 227–237. Edwards, C.R., Kipkie, W.B., Agar, G.E., 1980. The effect of slime coatings of the serpentine minerals, chrysotile and lizardite, on pentlandite flotation. International Journal of Mineral Processing 7, 33–42. Farrokhpay, S., Ametov, I., Grano, S., 2011. Improving the recovery of low grade coarse composite particles in porphyry copper ores. Advanced Powder Technology 22, 464–470. Gao, M., Forssberg, E., 1993. The influence of slurry rheology on ultra-fine grinding in a stirred ball mill. In: 18th International Mineral Processing Congress. Sydney, Australian. Johnson, S.B., Franks, G.V., Scales, P.J., Boger, D.V., Healy, T.W., 2000. Surface chemistry–rheology relationships in concentrated mineral suspensions. International Journal of Mineral Processing 58, 267–304.

98

D. Xu et al. / Minerals Engineering 39 (2012) 89–98

Kitano, T., Kataoka, T., Shirota, T., 1981. An empirical equation of the relative viscosity of polymer melts filled with various inorganic fillers. Rheologica Acta 20, 207–209. Muganda, S., Zanin, M., Grano, S.R., 2011. Influence of particle size and contact angle on the flotation of chalcopyrite in a laboratory batch flotation cell. International Journal of Mineral Processing 98, 150–162. O’Connor, C.T., Randall, E.W., Goodall, C.M., 1990. Measurement of the effects of physical and chemical variables on bubble size. International Journal of Mineral Processing 28, 139–149. Pashley, R.M., Kitchener, J.A., 1979. Surface forces in adsorbed multilayers of water on quartz. Journal of Colloid and Interface Science 71, 491–500. Prestidge, C.A., 1997a. Rheological investigations of galena particle interactions. Colloids and Surfaces A: Physicochemical and Engineering Aspects 126, 75–83.

Prestidge, C.A., 1997b. Rheological investigations of ultrafine galena particle slurries under flotation-related conditions. International Journal of Mineral Processing 51, 241–254. Pyke, B., Fornasiero, D., Ralston, J., 2003. Bubble particle heterocoagulation under turbulent conditions. Journal of Colloid and Interface Science 265, 141–151. Tangsathitkulchai, C., 2003. The effect of slurry rheology on fine grinding in a laboratory ball mill. International Journal of Mineral Processing 69, 29–47. Wang, Y., Forssberg, E., 1995. Dispersants in stirred ball mill grinding. Kona 13, 67– 77. Xu, D., Ametov, I., Grano, S.R., submitted for publication. Effect of medium viscosity on coarse particle flotation. International Journal of Mineral Processing. Yang, H.-G., Li, C.-Z., Gu, H.-C., Fang, T.-N., 2001. Rheological behavior of titanium dioxide suspensions. Journal of Colloid and Interface Science 236, 96–103.