pyrite mixtures

Powder Technology 267 (2014) 61–67

Contents lists available at ScienceDirect

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

Effect of regrinding and pulp aeration on the flotation of chalcopyrite in chalcopyrite/pyrite mixtures Clement Owusu ⁎, Daniel Fornasiero, Jonas Addai-Mensah, Massimiliano Zanin Ian Wark Research Institute, The ARC Special Research Centre for Particle and Material Interfaces, University of South Australia, Mawson Lakes, Adelaide, SA 5095, Australia

a r t i c l e

i n f o

Article history: Received 3 April 2014 Received in revised form 13 June 2014 Accepted 15 June 2014 Available online 22 June 2014 Keywords: Chalcopyrite Pyrite Flotation Regrinding Dissolved oxygen

a b s t r a c t In flotation, regrinding is used to increase the liberation of value minerals, such as chalcopyrite, from gangue minerals. As a result of the significant decrease in particle size following regrinding, flotation efficiency is generally worst. In this study it is shown that the flotation recovery (and rate) of chalcopyrite in chalcopyrite–pyrite mineral mixtures, which is lower after regrinding, could be partially restored with aeration. The increase in Cp flotation upon aeration was attributed to the creation of a more favourable environment for xanthate adsorption and dixanthogen formation whereas the decrease in Cp flotation after prolonged aeration is due to its increased oxidation as a result of galvanic interactions with pyrite upon regrinding. The amount of aeration required for maximum flotation decreased monotonically with increasing feed grind particle size. To rationalize the observed chalcopyrite flotation trends, a simple model which involves competitive adsorption of xanthate and oxygen on chalcopyrite and pyrite surface sites and underpinning oxygen consumption and mineral oxidation effect is proposed. © 2014 Elsevier B.V. All rights reserved.

1. Introduction As a result of the discovery of new ore bodies with declining grades that require fine grinding for increased liberation and efficient mineral separation, the flotation of fine particles has become increasingly important in recent times. At present, the greatest challenge which confronts flotation separation in the mineral industry is related to decreased efficiency in processing fine particles. Generally, fine hydrophobic particles, b 10 μm in size, display low flotation rate and poor recovery at conventional residence time in plan flotation machines [1,2]. The poor flotation behaviour of hydrophobic particles b 10 μm has been attributed to their low collision efficiency with rising bubbles due to their low mass/momentum [3]. Increasing the degree of turbulence/impeller speed and/or residence time during flotation has been found to increase the recovery of fine particles [4–6]. Fine grinding is however often required in practice. The key driving force for fine grinding is enhanced liberation of values and gangue minerals. The creation of new surfaces during fine grinding however, leads to the exposure of more active sites to alterations of the particle surface chemical properties which may influence significantly the flotation pulp chemistry (e.g. dissolved oxygen concentration and pulp oxidation potential). This may, in turn, significantly affect collector adsorption and oxidation,

⁎ Corresponding author. Tel.: +61 8 83023714; fax: +61 8 8302 3683. E-mail address: [email protected] (C. Owusu).

http://dx.doi.org/10.1016/j.powtec.2014.06.026 0032-5910/© 2014 Elsevier B.V. All rights reserved.

ultimately controlling the mineral particle flotation performance [7,8]. It is well known that a minimum amount of oxygen in the pulp (reflecting low Eh) is required for xanthate collectors to adsorb and, therefore, promote sulphide mineral flotation while an excess amount of oxygen (high Eh) leads to marked surface oxidation which has a detrimental effect [9–11]. Furthermore, galvanic interactions between sulphide mineral particles and the grinding media, with different electrocatalytic abilities for the reduction of oxygen, may induce changes in pulp and mineral surface chemistry which can affect the flotation performance [12–14]. Due to the complexity of copper ores, in which chalcopyrite and chalcocite are often finely interlocked with pyrite, to improve the value minerals liberation and copper flotation recovery and grade, regrinding to very fine particle size (sometimes b20 or 15 μm as practiced at the Prominent Hill copper mine and Telfer) is commonly performed. However, in most cases an increase in valuable mineral recovery is not observed upon regrinding, or at least not to the expected extent, suggesting that particle size and liberation are not the only key influential factors in determining the flotation performance [15]. This paper particularly investigates the effect of regrinding and aeration on the flotation performance of chalcopyrite and pyrite mineral mixtures. To rationalize the flotation behaviour, the amount of air/ oxygen reacting with chalcopyrite and pyrite and the resulting particle products formed is measured. Finally, a simple reaction mechanism which takes into account oxygen and xanthate concentrations and mineral surface area is proposed to qualitatively describe the observed chalcopyrite flotation.

62

C. Owusu et al. / Powder Technology 267 (2014) 61–67 100

2. Experimental

Single minerals of chalcopyrite (Cp) and pyrite (Py) obtained from Mannum Minerals and Peruvian Mine, respectively, were used for this study. The details of the chemical compositions of the minerals are reported in Table 1. Complementary XRD analysis showed that the chalcopyrite sample contained 6.3 wt.% pyrite while the pyrite contained 10% impurities of virtually no copper sulphide. A two stage grinding procedure (primary grinding to d80 of 105 μm and secondary grinding to d80 of 10, 20, and 30 μm) was used to produce the feed for this study. Fig. 1 shows the particle size distribution curves for the 10, 20, 30 and 105 μm grind size. The primary grinding was achieved in a closed Galigher laboratory mill with stainless steel grinding media while that of the secondary grinding was achieved in an attritioning mill (IsaMill) with ceramic beads (Ø3.5 mm) (Xstrata Technology, Australia) as the grinding media. Sodium isopropyl xanthate (SIPX) collector and polypropylene glycol (Dowfroth 250; AR grade) frother were used, both obtained from Cytec (Stamford, USA). Analytical grade lime and demineralised water were used throughout the experiment for pH control. 2.2. Mineral grinding and flotation In each test, 200 g of mineral (Cp alone or 100 g each of Cp and Py in mixtures) was added to 300 cm3 of demineralised water and known amount of lime in a Galigher laboratory mill to produce a pulp with a particle size distribution (PSD) of 80% passing 105 μm at pH 10. The mill product was then conditioned in a 1.5 dm3 flotation cell (Agitair LA-500R) at pH 10.5 with xanthate collector (200 g SIPX per ton of Cp in the feed) for 2 min. The conditioned pulp was reground in a 1 dm3 laboratory IsaMill (attrition mill) to produce a PSD of 80% passing 30, 20 or 10 μm. Product from the attrition mill was transferred back to the 1.5 dm3 flotation cell and conditioned without or with air (air flow rate = 3 dm3/min) for different times. Known amounts of SIPX additional collector (Table 2) and 25 g/t Dowfroth frother were added and conditioned for 2 and 1 min, respectively. In order to ensure the same collector coverage, the collector dosage was scaled to the mineral (BET) specific [35] surface area (Table 2) after IsaMilling. Impeller speed and pH during conditioning and flotation in the cell were maintained at 1000 rpm and 10.5, respectively for all experiments. Four flotation concentrates were collected after cumulative times of 1, 3, 6 and 10 min at an air flow rate of 2.5 dm3/min. The flotation froth was scraped every 10 s. The dry masses of the four concentrates together with their tails were measured and the samples assayed for their elemental compositions. Flotation rate constant (k) and maximum flotation recovery (%Rmax) were calculated by fitting the cumulative recovery (%R) versus its corresponding flotation time (t) of each concentrate with a first order rate equation, %R = %Rmax(1 − e− kt). Mineral recovery by true flotation and entrainment was calculated using the method developed by Ross [16]. 2.3. Oxygen demand test A number of techniques are currently used to determine the oxygen demand of flotation pulps and are all based on the method described by Spira and Rosenblum [17]. The method generally requires aeration/

Cp Py

80

d80 60

40 10 m 20 m 30 m 105 m

20

0 1

10

100

1000

Particle size (m) Fig. 1. Particle size distribution for primary and secondary grinding.

oxygenation of a freshly ground pulp to a particular oxygen level, after which aeration/oxygenation is ceased and the decay in oxygen level monitored as a function of time. This method is adequate for the laboratory, although it does not replicate the conditions in most processing plants. This is because in most processing plants the pulp is not aerated/oxygenated to a particular oxygen value or content but rather it is oxygenated or aerated for a constant amount of time. Another fundamental problem with the method is the fact that it is open to the atmosphere. As a result, it is likely that some amount of oxygen is lost to the atmosphere rather than being consumed by the pulp. With the DO demand experiment used in this study, all these short falls have been addressed. 2.3.1. Procedure The details of the apparatus used for the oxygen demand test have been described previously [18]. In each experiment, the IsaMill discharge pulp (either Cp or Py alone or Cp/Py mixture) was transferred into an airtight container and stirred for 1 min to obtain a homogeneous pulp after which the initial readings for pulp potential, DO, and pH were recorded. The pulp pH was adjusted to 10.5 and then air purged at a rate of 3 dm3/min for 3 min, after which the air supply was cut off for 5 min. This cycle of air on and off was repeated several times until the oxygen concentration in the vessel reached equilibrium. The pulp pH, DO and pulp potential values were logged continuously at time intervals of 10 s. The oxygen demand rate constant (Kla) for Cp, Py and Cp/Py mixtures was calculated after the air supply was cut off for all the cycles (Eq. (1)). −ðKlatÞ

DO ¼ DOo  e

ð1Þ

where DO and DOo are the dissolved oxygen concentrations at time t and t = 0, respectively. Dissolved oxygen (DO) (YSI membrane electrode), potential (Pt–Ag/ AgCl electrode) and pH (glass combination) probes connected to a multi-metre (TPS 90-FLMV, TPS Pty. Ltd.) were used to measure the DO concentration, pulp potential and pH, respectively. Standard buffer solutions of pH 7 and pH 10 were used to calibrate the pH electrode. Table 2 Collector concentration after secondary grinding scaled to mineral (BET) specific surface area for the various grind size (50 wt.% Py).

Table 1 Chemical composition of Cp and Py samples used in the study. Mineral

Cummulative undersize (%)

2.1. Materials and reagents

Elements (mass %) Zn

Fe

S

Ca

Cu

Si

Pb

Mn

Mg

0.01 0.18

28.9 44.9

29.0 53.5

1.76 0.44

26.4 0.13

4.2 0.7

0.34 0.02

0.02 0.02

0.68 0.08

d80 (μm)

BET surface area (m2/g)

SIPX (g/t) added after secondary grinding

10 20 30

1.76 1.33 0.76

33.1 25.0 14.3

C. Owusu et al. / Powder Technology 267 (2014) 61–67

2.4. Surface analysis A Kratos Axis Ultra X-ray photoelectron spectrometer (XPS) with an Al Kα monochromated X-ray source (1486.8 eV) operated at 130 W was used to record the species present on the mineral surface. Blocks of freshly fractured pure minerals of Cp and Py were placed in the pulp in the Cp–Py mineral mixture experiment during conditioning and were removed prior to aeration after pH adjustment to 10.5 and after 48 and 80 min of aeration for XPS analysis. Blocks were washed with a pH 10.5 solution to remove any suspended particles and introduced immediately in the fore-vacuum of the XPS spectrometer.

3. Results and discussion

63

(5). This may explain why Kla decreases in Fig. 2 with further air addition. Indeed, XPS analysis of the mineral surface (Fig. 3) shows that the amount of surface iron oxide/hydroxide increases after exposing the minerals to air, especially at the pyrite surface. −

1=2O2 þ H2 O þ 2e ⇔2OH

−

ð2Þ

2CuFeS2 þ 17=2O2 þ 9H2 O⇔2CuðOHÞ2 þ 2FeðOHÞ3 þ 4H2 SO4

ð3Þ

2FeS2 þ 15=2O2 þ 7H2 O⇔2FeðOHÞ3 þ 4H2 SO4

ð4Þ

3þ

Fe

−

þ 3OH ⇔FeðOHÞ3 :

ð5Þ

3.1. Dissolved oxygen consumption by Cp, Py and Cp/Py mixtures

1.0

8

DO level (ppm)

100 wt.% Py 100 wt.% Cp 50 wt.% Py

-1

Oxygen consumption rate constant, Kla, (min )

Upon turning on the air, the DO in the pulp increases up to a maximum value and then decreases to a minimum value when the air is turned off, for each cycle of aeration (Fig. 2), the latter because of O2 consumption by the mineral(s) present in the air-tight container. Fig. 2 shows the Kla for Cp and Py minerals alone and when they are mixed together in a ratio of 1:1 as a function of total volume of air supplied to the pulp. The Kla of the pulp in each aeration cycle was calculated using Eq. (1). The total volume of air supplied was also determined by multiplying the air flow rate by the aeration time. The data generally show similar trends with Kla decreasing with increasing the total volume of air supplied to the pulp. The results in Fig. 2 also show that Py has high Kla than Cp, with more air needed for Py than for Cp to reach equilibrium, which is in agreement with the more cathodic nature of Py than Cp [19]. The Kla obtained when the minerals are mixed together is not an average of Py and Cp. This may be attributed to synergistic effect and galvanic interactions occurring between the minerals in the mixtures. The high Kla values observed for the minerals at the initial state of air supply is attributed to the reduction of O2 at the mineral surface according to Eq. (2) and the oxidation of the sulphide minerals according to Eqs. (3) and (4) for Cp and Py, respectively. Of course, because Py is a more cathodic mineral, Py contributes more to reaction (2) and much less to reaction (4) while the opposite is true for Cp. Furthermore, the hydroxide ions produced in reaction (2) react with iron at the Py surface to form ferric hydroxide as in Eq. (5) [20]. The layer of these iron hydroxide and sulphate species formed at the mineral surface constitutes a barrier which slows down the access of O2 to the mineral surface, and therefore decreases O2 reaction according to Eqs. (2) to

0.8

6 4 2 0

0.6

Cp Py

0

20

40

60

80

100

Aeration time (min)

0.4

0.2

0.0 0

20

40

60

80

100

120

140

160

180

Volume of air supplied (dm 3) Fig. 2. Oxygen consumption rate constant, Kla, for Cp (▲) and Py (●) alone and Cp–Py mixture (◊) (50 wt.% Py) as a function of total volume of air supplied. The dissolved oxygen measured in the pulp for several cycles of air on and air off for Cp (filled symbols) and Py (empty symbols) alone is shown in the inset (d80 = 20 μm).

3.2. XPS results Fig. 3 shows the Fe2p XPS spectra of Cp and Py at different aeration times. It is clear from these results that the proportion of iron oxide/ hydroxide (broad band at 711–712 eV) [21], at the surface of the minerals increases with aeration time, especially for pyrite. Only one component at 932.0 eV is observed in the Cu2p XPS spectra (not shown); it is attributed to copper in a Cu(I) state [21]. No or little Cu(II) species such as copper oxide/hydroxide were observed (bands near 934 eV and 935 eV). Peaks for sulphate and other sulphoxy species were small in the S2p XPS spectra (not shown) and remain more or less constant with increasing aeration. The summary of XPS data is provided in Table 3. This XPS analysis indicates that more iron oxides/hydroxides are formed at the Cp and Py surfaces with increasing aeration time, and this increase in surface iron oxide/hydroxide is much larger for Py than for Cp in the aeration time investigated, which confirms the results of the DO experiment that more oxygen is consumed by Py than for Cp. 3.3. Flotation results Fig. 4 shows the flotation recovery of Cp as a function of volume of air supplied for the Cp (100 wt.% Cp) and Cp/Py (50/50) mixture after 10 min of flotation (d80 = 20 μm) after subtraction of entrainment. For Cp alone, its recovery decreases from 96% without aeration to 86% after addition of 240 dm3 of air to the flotation cell. For the Cp/Py mixture, Cp recovery is much less without aeration, 72%, but it increases up to a maximum value of 80% after the addition of 108 dm3 of air before decreasing with further air addition (71% at 240 dm3 of air). On the other hand, Py recovery after 10 min of flotation (results not shown) decreased continuously upon aeration (from 8% without air to 2% after 240 dm3 of air addition). This trend of Cp recovery with air addition obtained for the mineral mixture is reminiscent of previous flotation results of chalcopyrite, and other copper sulphide minerals, with increasing pulp oxidation potential (Eh) or air/oxygen concentration, whereas a maximum in flotation recovery was found at an intermediate pulp oxidation potential or air/oxygen concentration [9,10,14,22]. The increase in Cp recovery observed in Fig. 4 and the previous studies is attributed to xanthate adsorption and the formation of surface hydrophobic species of metal xanthate/dixanthogen while the decrease in Cp recovery is the result of Cp oxidation with the formation of the surface hydrophilic species of mainly iron oxide/hydroxide (Eqs. (3)–(5) and XPS results) at high pulp air/oxygen concentration or Eh values. From Fig. 4, it is apparent that there is enough air/oxygen in the pulp for xanthate adsorption and therefore maximum Cp recovery without the need for further air addition, which is not the case when Cp is mixed with Py. It is also interesting to point out that Cp recovery for the mineral mixture was lower and could not be restored with air addition to recoveries found for Cp alone. It is possible that there is not enough xanthate for Cp because it is consumed by Py; however Py

64

C. Owusu et al. / Powder Technology 267 (2014) 61–67

80 min FeO/OH

48 min 0 min

730

725

720

715

80 min

Relative Intensity, Arbitrary Units

Relative Intensity, Abitrary unit

FeS2

710

705

700

FeO/OH CuFeS2

48 min 0 min

730

725

720

715

710

705

700

Binding Energy (eV)

Binding Energy (eV)

Fig. 3. Fe2p XPS spectra of Py (left) and Cp (right) at different aeration times.

3.3.1. Effect of grind particle size (d80) on Cp recovery Fig. 5 shows the effect of aeration on Cp flotation recoveries after 10 min of flotation for Cp–Py mixture (50% Py) at three different grind particle sizes, d80 of 10, 20, and 30 μm after subtraction of entrainment. The results clearly showed that grind particle size has an important role in Cp flotation. Without aeration, Cp recovery decreases from 77% to 72% and 36% for d80 of 30, 20 and 10 μm, respectively. This was associated with a decreased in Cp flotation rate constant from 0.71 min−1 to 0.50 min−1 and 0.12 min−1 for d80 of 30 μm, 20 and 10 μm, respectively. With an increase in pulp air/oxygen content through aeration, the Cp flotation kinetics and recovery increased up to a maximum value before decreasing with further air/oxygen addition for the three grind particle sizes, the effect being more pronounced for the lowest grind size. It is also clear that more air is needed to reach maximum recovery with decreasing grind particle size, 30 dm3, 108 dm3 and 240 dm3 for grind sizes of 30, 20 and 10 μm, respectively. Py recovery on the other hand increased with decreasing grind particle size without aeration (Fig. 5). However, upon aeration, a continuous decrease of Py recovery was observed for all the particle grind size. Fig. 6 shows that Cp recovery (at 10 min of flotation) after optimum aeration (the aeration time that produced maximum Cp recovery in Fig. 5 after entrainment subtraction) increased from 36 to 52%, from 72 to 80% and from 77 to 79% for d80 of 10, 20 and 30 μm, respectively. This was associated with an increase in Cp flotation rate constant from 0.12 to 0.23 min− 1, from 0.64 to 0.78 min−1 and from 0.41 to 0.74 min− 1. On the other hand, Py recovery after optimum aeration decreased from 18 to 11%, from 8 to 5% and from 7 to 2%, respectively, for the d80 of 10, 20 and 30 μm (Fig. 6). Therefore, a significant

Table 3 Surface average weight percentage of elements on mineral surface for different Cp/Py mixtures. Aeration time

Cu Fe O C S

0 min

48 min

improvement in Cp separation from Py (increased Cp recovery together with decreased Py recovery) can be obtained with optimising aeration at the three grind particle sizes. However, Cp flotation recoveries and rate are reduced with grinding finer, especially below 20 μm. The observed increase in Cp recovery with increasing air/oxygen content is due to collector adsorption and formation of hydrophobic species of metal xanthate and dixanthogen while the decrease in Cp recovery results from the oxidation of the Cp surface [9,10,14]. As the mineral grind size becomes finer, the mineral surface area increases, which requires not only more xanthate (Table 2) but also more air/ oxygen to reach a maximum in Cp recovery, hence the shift of maximum Cp recovery to higher aeration volumes with decreasing grind size observed in Fig. 5. It is tempting to attribute the overall decrease in Cp recovery with decreasing grind size to the well-known lower collision efficiency of fine particles with gas bubbles [3,25,26]. However, it is found in Fig. 6 that Py recovery increases as their particles are ground finer. It is more plausible that this decrease in Cp recovery and increase in Py recovery may be the result of galvanic interaction between the two minerals which increase with decreasing grind size (with increasing mineral surface area), and results in increase copper dissolution and oxidation of Cp and therefore increased Cu activation of Py. The maximum in Cp recovery curves in Fig. 5 shifts to higher air/oxygen

100

95

90

Cp recovery (%)

recoveries were very low (b10%). The decrease in Cp recovery can be attributed to further oxidation of the Cp surface in the presence of Py due to galvanic interactions [13,23].

85

100 wt.% Cp 50 wt.% Py

80

75

70

80 min

Cp

Py

Cp

Py

Cp

Py

9.4 10.1 17.6 43.5 19.4

0.7 16.6 21.3 35.6 25.9

11.6 10.8 18.2 41.6 17.8

1.0 11.4 31.9 39.6 16.1

6.3 7.8 29.9 42.6 13.4

0.7 12.4 38.2 36.5 12.3

65 0

50

100

150

200 3

Total volume of air supplied (dm ) Fig. 4. Cp recovery after 10 min of flotation as a function of volume of air supplied for Cp alone and 50 wt.% Py mineral mixture (d80 = 20 μm in secondary grinding; pH = 10.5). Adapted from [24].

C. Owusu et al. / Powder Technology 267 (2014) 61–67

65

300

100

3

Optimum volume of air (dm )

250

90 3

Optimum volume of air (dm )

250

Cp recovery (%)

80

70 10 m 20 m 30 m

60

50

200

150

200

150

100

50

0 0.6

100

0.8

1.0

1.2

1.4

1.6

1.8

2

BET area (m /g)

50 40

0

30 0

100

200

10

300

20

30

40

50

Grind particle size (d80, m)

3

Total air supplied (dm ) Fig. 5. Cp recovery (after 10 min of flotation) as a function of volume of air supplied for three grind particle sizes, d80 (50 wt.% Py) [24].

Fig. 7. Correlation between the optimum volume of air for maximum Cp recovery against grind particle size (d80) and particle surface area (inset) [24].

concentrations as grind particle size decreases, suggesting that more oxygen is required upon fine grinding. A plot in Fig. 7 of the optimum amount of air required for maximum Cp recovery against grind particle size shows a linear correlation. Since surface area is the main changing parameter as one grinds finer, this optimum amount of air for maximum Cp recovery also correlated well with the mineral (BET) specific surface area [24]. These suggest that the optimum amount of air required for good mineral separation is greatly influenced by the mineral surface area.

equilibrium constant for that reaction. Oxygen also reacts with the Cp surface resulting in Cp oxidation as in Eq. (3). Eqs. (7) and (8) may simply represent this oxidation reaction.

3.4. The reaction mechanisms of xanthate and oxygen on Cp surface for different grind particle sizes It is well know that the adsorption of collector (xanthate) on the surface of value metal sulphide minerals occurs via an electrochemical reaction which is coupled with the reduction of oxygen (Eq. (1)) on the mineral surface [12,27–29]. Therefore, the adsorption of xanthate (X) on to chalcopyrite may then be simply represented by Eq. (6) Ks

Cp þ O2 þ X ⇔ CpO2 X

ð6Þ

where Cp is the surface sites, CpO2X is the adsorbed xanthate species such as metal xanthate and/or dixanthogen [30,31] and K1 is the

Ks

Cp þ O2 ⇔ CpO2 Ks

CpO2 þ O2 ⇔ CpðO2 Þ2

ð8Þ

where CpO2 and Cp(O2)2 are surface copper and iron oxide/hydroxide and/or sulphoxy species [32], and K2 is the equilibrium constant for these reactions (Eq. (8) indicates that more than one monolayer of oxygen molecules can be adsorbed on Cp). Of course, xanthate competes with oxygen for the Cp surface sites, and the different proportions of surface xanthate species and metal oxide/hydroxide and sulphate species are regulated by the relative values of K1 and K2. The total concentrations of Cp surface sites, xanthate molecules and oxygen molecules are respectively:   ½CpT ¼ ½Cp þ ½CpO2 X þ ½CpO2  þ CpðO2 Þ2 ½XT ¼ ½X þ ½CpO2 XÞ

100

100 m (No air) m (Opt. air) m (No air) m (Opt. air) m (No air) m (Opt. air)

80

% Cp species

10 10 20 20 30 30

80

Cum. Cp recovery (%)

ð7Þ

60

40

[CpO2X]

[Cp]

[Cp(O2)2]

60

40

[CpO2] 20

20

0 -2.6

0 0

2

4

6

8

10

12

14

16

18

20

-2.4

-2.2

-2.0

-1.8

-1.6

-1.4

log[O2]

Cum. Py recovery (%) Fig. 6. Cp recovery versus Py recovery without and with optimum aeration at different flotation times and grind particle sizes, d80 (50 wt.% Py).

Fig. 8. Calculations of the proportion of species at the chalcopyrite surface as a function of the total oxygen concentration (logK1 = 9; logK2 = 6.5). Note that the concentration of CpO2X has been multiplied by 100 for a clearer comparison with the other Cp species.

66

C. Owusu et al. / Powder Technology 267 (2014) 61–67

  ½O2 T ¼ ½O2  þ ½CpO2 X þ ½CpO2  þ 2 CpðO2 Þ2 : Fig. 8 shows an example of the calculations (by matrix inversion) of these surface species (Cp, CpO2X, CpO2 and Cp(O2)2) when the oxygen concentration was increased from 10–2.6 M to 10–1.5 M for values of K1 and K2 of 109 M−2 and 106.5 M−1, respectively. The total concentrations of surface sites ([Cp]T) and xanthate ([X]T) were calculated as 1.50–2 and 2.10–4, respectively. As expected from Eqs. (6) to (8), the proportion of free Cp surface site decreases with increasing oxygen concentration while that of CpO2X increases up to a maximum value where all the xanthate molecules have been adsorbed (2.10–4 M xanthate and 10– 1.8 M oxygen). With a further increase in oxygen concentration, xanthate at the surface is replaced by oxygen to form the Cp(O2)2 species which concentration increases. Of course the adsorbed xanthate species (CpO2X) are hydrophobic while the Cp sites (Cp) and metal oxide/hydroxide and sulphate species (CpO2 and Cp(O2)2) are hydrophilic. Therefore, the proportion of adsorbed xanthate species (CpO2X) can represent Cp recovery which also increases with increasing aeration up to a maximum value before decreasing as seen in Fig. 4 and in other studies [10,14] with increasing aeration or Eh value. Because Cp recovery was much lower in the absence of xanthate in the experimental conditions of this study and for simplicity in the calculations, the hydrophobic species of polysulphide and elemental sulphur which have been shown to exist on the Cp surface and be responsible for Cp flotation in the absence of collector were not considered in this study [9,21,33]. Similar equations are proposed for the reaction of oxygen with the pyrite surface (Eq. (9)). Ks

Py þ O2 ⇔ PyO2

Ks

Ks

PyO2 þ O2 ⇔ PyðO2 Þ2 PyðO2 Þ2 þ O2 ⇔ PyðO2 Þ3

Ks

PyðO2 Þ3 þ O2 ⇔ PyðO2 Þ4

Ks

PyðO2 Þ4 þ O2 ⇔ PyðO2 Þ5

ð9Þ

where Py(O2)n represents iron oxide/hydroxide and sulfoxy species [34] and K3 is the equilibrium constant for these reactions. More oxygen reactions at the Py surface than at the Cp surface were considered in the calculations to reflect the observation that more oxygen is consumed by Py than Cp (Fig. 2). Because the recoveries of Py were much lower than those of Cp (Fig. 6) and for simplicity in the calculations the adsorption of xanthate onto Py was not taken into account. By knowing all the equilibrium constants in Eqs. (6)–(9), the total concentrations of Cp and Py surface sites, and xanthate added, the concentration of each species in Eqs. (6)–(9) can be calculated by matrix

inversion as a function of the concentration of oxygen added. The results of these calculations using Eqs. (6)–(9), K1, K2 and K3 values of 109 M−2, 106.5 M−1 and 107 M−1, respectively are shown in Fig. 9. For the calculations for the mineral mixture with 50 wt.% Py, the concentrations of Cp surface sites and xanthate were halved (1.5–2 and 1.05–4 M for 100 wt.% Cp and 50 wt.% Py, respectively) to replicate the experimental conditions of Fig. 3, knowing that Cp and Py have approximately the same surface area, 2.4 and 2.6 m2/g, respectively. The results of calculations show that first the proportion of CpO2X on the Cp surface increases with an increasing oxygen concentration up to a maximum value before decreasing and secondly, the maximum in %CpO2X shifts to higher oxygen concentrations in the presence of Py, which is consistent with the trends of Cp recovery observed in Fig. 4 and which in turn validates the choice of equations used and assumptions made in these calculations. At low oxygen concentrations, the oxygen is preferentially consumed by Py (Fig. 2) and therefore, more oxygen is required to reach a maximum value in the proportion of CpO2X, hence the shift in the position of the maximum. To account for the effect that particle grind size has on Cp flotation, similar calculations were made using the same equations (i.e. Eqs. (6)–(9)) and the same values of the equilibrium constants, K1 = 109 M−2, K2 = 106.5 M−1 and K3 = 107 M−1. The concentration of xanthate was scaled to the mineral specific surface area (Table 2) to mimic the experimental conditions of Fig. 5: 0.96–4, 1.05–4 and 1.12–4 M for d80 of 30, 20 and 10 μm, respectively. Calculation results in Fig. 9 show that the maximum in the proportion of the hydrophobic species CpO2X at the Cp surface shifts to higher oxygen concentrations as the particle grind size is decreased from 30 to 20 and 10 μm, as it was the case for the maximum in Cp recoveries observed in Fig. 5. This shift to higher oxygen concentrations is explained in Eq. (6) where, if the number of Cp sites is increased (because the mineral surface area increases), more oxygen is required for all the xanthate molecules to adsorb. XPS results have shown that the Cp surface is more oxidised/covered with more iron oxide/hydroxide with increasing aeration, especially at long aeration times (increase in surface oxygen proportion from 31% to 52% between 48 and 80 min of aeration). Therefore, oxidation of the Cp surface may be an explanation for this large decrease in Cp recovery observed in Fig. 5. To simulate this, the value of K2 in Eqs. (7) and (8) (which represent the reaction of oxygen with Cp leading to Cp oxidation) was increased from 106.5 to 107 M−1. Fig. 10 shows that with an increase in K2 value, the maximum in CpO2X does indeed decrease as less xanthate can now adsorb because of the increased Cp oxidation.

1.0

0.5

0% Py 10 µm

0.4

0.6

% CpO2X

% CpO2X

0.8

0.4

20 µm

0.3

30 µm

0.2

50% Py

0.0 -2.2

10 µm logK2=7

0.1

0.2

-2.0

-1.8

-1.6

-1.4

-1.2

log[O2] Fig. 9. Calculations of the proportion of CpO2X species at the chalcopyrite surface as a function of the total oxygen concentration at different percentages of pyrite in the Cp–Py mineral mixture (logK1 = 9; logK2 = 6.5; logK3 = 7).

0.0 -2.2

-2.0

-1.8

-1.6

-1.4

-1.2

log[O2] Fig. 10. Calculations of the proportion of CpO2X species at the chalcopyrite surface as a function of the total oxygen concentration at different grind particle sizes, d80 for 50% Py in mixture (logK1 = 9; logK2 = 6.5; logK3 = 7, except for curve 10 µm where logK2 = 7).

C. Owusu et al. / Powder Technology 267 (2014) 61–67

4. Conclusions These chalcopyrite–pyrite mineral flotation studies show that the decrease in particle size upon regrinding leads to a marked decrease in chalcopyrite flotation recovery and rate, notwithstanding the fact that the concentration of xanthate collector was scaled to the mineral surface area. Chalcopyrite recovery could be partially restored to a maximum value via pulp aeration to a certain point beyond which further aeration was detrimental. Aeration of the pulp improved the selective separation of chalcopyrite from pyrite especially for the smaller particle size distributions. The decrease in chalcopyrite flotation was mostly attributed to increased chalcopyrite oxidation as a result of increasing galvanic interaction with pyrite as the mineral surface area increases with further regrinding. Pyrite flotation remained low and further decreased with aeration. The amount of aeration required for maximum flotation correlated negatively to the feed grind particle size. Thus, more air/oxygen is required as particle size decreases. A simple model was proposed to explain the trends in chalcopyrite flotation. The model mainly relies on the competitive adsorption of xanthate and oxygen on chalcopyrite and pyrite surface sites which number increases with further regrinding, and the much larger oxygen consumption by pyrite than by chalcopyrite. Acknowledgement Financial support for this work from the AMIRA International, P260F project, University of South Australia and the Ghanaian Government is gratefully acknowledged. Useful discussions and comments provided by Professor W. Skinner (UniSA) and Michael Young (Xstrata Tech.) are appreciated. References [1] W. Trahar, L. Warren, The flotability of very fine particles — a review, Int. J. Miner. Process. 3 (2) (1976) 103–131. [2] W. Trahar, A rational interpretation of the role of particle size in flotation, Int. J. Miner. Process. 8 (4) (1981) 289–327. [3] Z. Dai, D. Fornasiero, J. Ralston, Particle–bubble collision models — a review, Adv. Colloid Interf. Sci. 85 (2) (2000) 231–256. [4] J. Pease, D. Curry, M. Young, Designing flotation circuits for high fines recovery, Miner. Eng. 19 (6) (2006) 831–840. [5] S. Grano, Effect of impeller rotational speed on the size dependent flotation rate of galena in full scale plant cells, Miner. Eng. 19 (13) (2006) 1307–1318. [6] J. Ralston, D. Fornasiero, S. Grano, J. Duan, T. Akroyd, Reducing uncertainty in mineral flotation — flotation rate constant prediction for particles in an operating plant ore, Int. J. Miner. Process. 84 (1) (2007) 89–98. [7] C. Greet, G. Small, P. Steinier, S. Grano, The Magotteaux Mill: investigating the effect of grinding media on pulp chemistry and flotation performance, Miner. Eng. 17 (7) (2004) 891–896. [8] S. Grano, The critical importance of the grinding environment on fine particle recovery in flotation, Miner. Eng. 22 (4) (2009) 386–394. [9] G. Heyes, W. Trahar, Oxidation–reduction effects in the flotation of chalcocite and cuprite, Int. J. Miner. Process. 6 (3) (1979) 229–252. [10] P. Richardson, G. Walker, The flotation of chalcocite, bornite, chalcopyrite, and pyrite in an electrochemical-flotation cell, 15th International Mineral Processing Congress (15th Congres International de Mineralurgie)1985. 198–210.

67

[11] Z. Kristall, S.R. Grano, K. Reynolds, R.S.C. Smart, J. Ralston, An Investigation of Sphalerite Flotation in the Murchison Zinc Concentrator. 1994, Australasian Institute of Mining and Metallurgy, 1994. [12] S. Rao, J. Finch, Galvanic interaction studies on sulphide minerals, Can. Metall. Q. 27 (4) (1988) 253–259. [13] Y.J. Peng, S. Grano, D. Fornasiero, J. Ralston, Control of grinding conditions in the flotation of chalcopyrite and its separation from pyrite, Int. J. Miner. Process. 69 (1–4) (2003 Mar) 87–100. [14] S. He, W. Skinner, D. Fornasiero, Effect of oxidation potential and zinc sulphate on the separation of chalcopyrite from pyrite, Int. J. Miner. Process. 80 (2) (2006) 169–176. [15] J. Frew, K. Davey, R. Glen, Smart RSC. Effects of fine grinding on flotation performance: zinc regrind at Cominco Alaska's Red Dog Mine, Proc Fifth Mill Operators' Conference (Aust Inst Min Metall, Carlton 1994)1994. 1994 (p. 287-8). [16] V. Ross, Comparison of methods for evaluation of true flotation and entrainment, Trans. Inst. Min. Metall. Sect. C Miner. Process. Extr. Metall. 100 (1991) C121–C126. [17] P. Spira, F. Rosenblum, The oxygen demand of flotation pulps, 6th Annual Meeting of the Canadian Mineral Processors CIM, Ottawa; 19741974. 74–106. [18] C. Owusu, J. Addai-Mensah, D. Fornasiero, M. Zanin, Estimating the electrochemical reactivity of pyrite ores — their impact on pulp chemistry and chalcopyrite flotation behaviour, Adv. Powder Technol. 24 (2013) 801–809. [19] M.Q. Xu, J.A. Finch, S.R. Rao, D. Lin, Reverse flotation of pyrite from zinc-concentrate using nitrogen, Miner. Eng. 8 (10) (1995 Oct) 1159–1173. [20] D. Fornasiero, V. Eijt, J. Ralston, An electrokinetic study of pyrite oxidation, Colloids Surf. 62 (1) (1992) 63–73. [21] G. Fairthorne, D. Fornasiero, J. Ralston, Effect of oxidation on the collectorless flotation of chalcopyrite, Int. J. Miner. Process. 49 (1–2) (1997) 31–48. [22] X.-M. Yuan, B. Pålsson, K. Forssberg, Flotation of a complex sulphide ore I. CuZn selectivity control by adjusting pulp potential with different gases, Int. J. Miner. Process. 46 (3) (1996) 155–179. [23] C. Owusu, e Brito, S. Abreu, W. Skinner, J. Addai-Mensah, M. Zanin, The influence of pyrite content on the flotation of chalcopyrite/pyrite mixtures, Miner. Eng. 55 (2014) 87–95. [24] C. Owusu, J. Addai-Mensah, M. Zanin, Correlations between optimum aeration and pyrite content in a modelled chalcopyrite–pyrite xanthate flotation system, Chemeca 2013 Conference, Brisbane; 20132013. [25] M. Weber, D. Paddock, Interceptional and gravitational collision efficiencies for single collectors at intermediate Reynolds numbers, J. Colloid Interface Sci. 94 (2) (1983) 328–335. [26] H. Schulze, B. Radoev, T. Geidel, H. Stechemesser, E. Töpfer, Investigations of the collision process between particles and gas bubbles in flotation—a theoretical analysis, Int. J. Miner. Process. 27 (3) (1989) 263–278. [27] S.A. Allison, L.A. Goold, M.J. Nicol, A. Granville, A determination of the products of reaction between various sulphide minerals and aqueous xanthate solutions and a correction of the products with electrode rest potentials, Metall. Trans. 3 (1972) 2613–2617. [28] R. Woods, Electrochemistry of sulphide flotation, Princ. Miner. Flotation (1984) 91–115. [29] D.W. Fuerstenau, R.H. Urbina, Flotation fundamentals, Reagents in Mineral Technology, Marcel Dekker, New York, 1987. 1–38. [30] J. Leppinen, C. Basilio, R. Yoon, In-situ FTIR study of ethyl xanthate adsorption on sulfide minerals under conditions of controlled potential, Int. J. Miner. Process. 26 (3) (1989) 259–274. [31] G. Senior, W. Trahar, The influence of metal hydroxides and collector on the flotation of chalcopyrite, Int. J. Miner. Process. 33 (1) (1991) 321–341. [32] A.N. Buckley, R. Woods, An X-ray photoelectron spectroscopic study of the oxidation of chalcopyrite, Aust. J. Chem. 37 (12) (1984) 2403–2413. [33] J.B. Zachwieja, J.J. McCarron, G.W. Walker, A.N. Buckley, Correlation between the surface composition and collectorless flotation of chalcopyrite, J. Colloid Interface Sci. 132 (2) (1989) 462–468. [34] A. Buckley, R. Woods, The surface oxidation of pyrite, Appl. Surf. Sci. 27 (4) (1987) 437–452. [35] S. Brunauer, P.H. Emmett, E. Teller, Adsorption of gases in multimolecular layers, J. Am. Chem. Soc. 60 (2) (1938) 309–319.