Critical Coalescence Concentration

Figure 12.21 shows the common trend observed for all frothers: a rapid initial decrease in size to a transition concentration where the minimum D₃₂ is reached. Cho and Laskowski (2002) introduced the term critical coalescence concentration (CCC) to refer to this transition concentration. Their graphical method of estimating CCC is included on Figure 12.21. Nesset et al. (2007) used a three-parameter model to fit the data and estimated CCC as CCC95, the concentration giving 95% reduction in bubble size compared to water alone.

With due precaution the two procedures give similar estimates of CCC (Finch and Zhang, 2014). Table 12.4 gives the CCC for some commercial frothers determined in air–water using a 0.8 m³ mechanical cell at a superficial gas (air) velocity of 0.5 cm s⁻¹. From the table, the CCC ranges from 5 ppm (PFW31) to 23 ppm (NF240); in terms of bubble size reduction, these two frothers can be classed as “strong” and “weak.” At concentrations above CCC all frothers give similar minimum bubble size, although some dependence on frother type is evident (Finch and Zhang, 2014). Using a large cell (0.8 m³ in the case of Figure 12.21) rather than a laboratory-size unit is recommended for estimating CCC (Zhang et al., 2009).

As all frothers yield a similar D₃₂-C trend, data for different frothers can be reduced to a common trend by plotting D₃₂ against C/CCC. Knowing the frother CCC bubble size at any concentration can be predicted from the D₃₂ trend with C/CCC.

The CCC will depend to an extent on factors such as gas rate and perhaps method of bubble generation. Bubble size will increase as gas rate increases, reflecting that the energy per unit mass of air imparted by the gas dispersion device has decreased corresponding to less break-up of the air. The increase in Sauter mean diameter with air rate has been modeled by Nesset et al. (2012), but as an approximation D₃₂ increase as J_g^0.25 (Xu et al., 1991). In Figure 12.21, this means as gas rate is increased the curve shifts upward and to the right; that is, the CCC increases with increasing gas rate. The choice of J_g=0.5 cm s⁻¹ in Table 12.4 is to try to standardize reporting of CCC.

There is a lack of data to gauge the effect on bubble size of the air dispersion device. The general D₃₂-C trend will not be changed but may be shifted to higher or lower C values compared to Figure 12.21, and thus CCC would be influenced by flotation machine type. For the present, the data in Table 12.4 provide a useful guide. The presence of solids does not appear to have a significant impact on bubble size or CCC (Nesset et al., 2012; Grandon and Alvarez, 2014), although an effect can be anticipated if a combination of fine particles and high solids content increase slurry viscosity significantly.

CCC and Practical Implications

The CCC concept indicates it is frother concentration in solution that controls bubble size. Operating practice is to quote frother dosage based on solids feed rate (kg t⁻¹). Useful for accounting purposes, it is not suited to understanding the role of frother. Table 12.5 calculates frother concentration in solution from published data (Damjanovic´ and Goode, 2000) showing concentrations appear to be above CCC at these operations. To maximize the benefit of frother addition a concentration slightly above CCC is the target; a concentration lower than CCC means (from Figure 12.21) any change in concentration has a large impact on bubble size which impacts kinetics. Increasing C much beyond CCC, apart from cost, risks excess water being carried to the froth, increasing entrainment (Zhang et al., 2010; Welsby, 2014). In the case of coal and other carbonaceous matter, the addition of frother may appear to greatly exceed CCC but this is because some frother is adsorbed by the solids; what remains in solution is the few ppm enough to produce the bubble size reduction (Gredelj et al., 2009). Some solids may adsorb one type of frother but not another. An example is talc, which adsorbs polyglycol frothers but not alcohols (Kuan and Finch, 2010).

Nesset et al (2012) incorporated CCC in a model to predict bubble size in mechanical flotation cells as a benchmark for operations: comparing predicted against measured D₃₂ could indicate potential for further decreases in bubble size (see Section 12.14.2). Plants operating below CCC may find that increasing frother dosage is not compatible with froth stability requirements, giving excess water recovery, for example. If this is the situation it may suggest that a different frother should be considered.

Copper Ions

Activation in sulfide systems is reviewed by Finkelstein (1997). The prime example is activation by copper ions. Used for a variety of sulfides (Allison and O’Connor, 2011), the classic use remains the activation of sphalerite. Sphalerite is not readily floated by a xanthate collector, for two reasons: (1) the collector product zinc xanthate is, compared to other base metal xanthates, relatively soluble and does not provide stable hydrophobic sites; and (2) because sphalerite is a poor semiconductor it does not readily support electron transfer reactions (Section 12.4.4). Floatability can be improved by the use of large quantities of long-chain xanthates, but a more efficient method is to use copper sulfate as an activator, which is readily soluble and dissociates into copper ions in solution. Activation is due to the exchange of Cu for Zn in the surface lattice of sphalerite:

$\mathrm{ZnS}+{\mathrm{Cu}}^{2+}\leftrightarrow \mathrm{CuS}+{\mathrm{Zn}}^{2+}$ (12.9)

(12.9)

This exchange is favored (a) because of their relative position in the electrochemical series (copper is above zinc and thus when together zinc will be oxidized to zinc ions and copper ions reduced to copper), and (b) because the sphalerite lattice can accommodate the copper ions without undue distortion. The exchange mechanism is evidenced by the fact that a mole of Cu lost from solution is replaced by a mole of Zn in solution (Fuerstenau et al., 2007). Exchange continues till the capacity to hold Cu in the sphalerite surface is reached. The addition rate in practice is anywhere between 0.2 and 1.6 g Cu per kg Zn (Damjanović and Goode, 2000; Finch et al., 2007a). That copper is held in the lattice, rather than more loosely on the surface by a physical adsorption mechanism, is confirmed by copper not being extracted by ethylene diamine tetra-acetic acid (EDTA) (Kant et al., 1994). The sphalerite acts as a “sink” for copper ions.

Sphalerites almost always contain variable iron in solid solution, the low Fe variety being pale brown and the high Fe end member, marmatite (with up to 10% Fe), being black. The presence of iron is recognized in the commonly quoted formula (Zn,Fe)S. In general, sphalerite flotation is not significantly affected by the presence of iron, although it is reported that misplacement to lead concentrates may be greater for the low iron sphalerite variety (Zieliński et al., 2000).

The presence of the copper in the sphalerite surface has two effects: (1) with xanthate the copper forms highly insoluble copper xanthate; and (2) surface conductivity increases to support the electrochemical xanthate adsorption mechanism. The anodic reaction of xanthate also appears to be associated with Cu²⁺ acting as an electron acceptor and reducing to Cu⁺ as the reaction product identified is Cu(I)-xanthate (Wang et al., 1989a,b).

The main use of copper sulfate as an activator is in the differential flotation of lead–zinc and copper–zinc ores, where after lead or copper flotation the sphalerite is activated and floated. Selectivity in the zinc flotation stage is generally against iron sulfides, pyrite, and pyrrhotite. The common approach is to first raise pH to depress the iron sulfides then add copper sulfate. The addition of copper sulfate itself, however, can promote selectivity by reducing pyrite flotation (Dichmann and Finch, 2001; Boulton et al., 2003). Experiments on sphalerite/pyrite/xanthate systems showed that in the absence of copper both sphalerite and pyrite floated, but addition of copper altered the balance of competition for xanthate in favor of the sphalerite and the pyrite became “starved” of xanthate. While the selectivity mechanism remains the subject of investigation (Chandra and Gerson, 2009), a practical outcome is in the order of reagent addition: rather than the common practice of first raising pH, adding copper sulfate first may both activate sphalerite and “starve” the pyrite of xanthate, to be followed by raising pH to complete depression of the pyrite. This CuSO₄–lime sequence of reagent addition has shown increased sphalerite/pyrite selectivity over the lime–CuSO₄ sequence in plant practice (Finch et al., 2007a). Provided copper addition rate is kept below the capacity of sphalerite to hold the copper, there appears to be little activation of the pyrite.

Copper ions are sometimes connected to inadvertent or accidental activation, the unintentional activation of other minerals, which hampers selectivity. Release of copper ions into solution from secondary copper minerals such as chalcocite (Cu₂S) can be an uncontrolled source of accidental sphalerite activation, sometimes with overwhelming effects on Cu/Zn selectivity (McTavish, 1985). On occasion, accidental activation of non-sulfide gangue by Cu ions is also suspected.

Other metals ions can cause accidental activation and hamper selectivity, notably Pb in Pb/Zn differential flotation (see Section 12.8) and possibly Ni in pentlandite/pyrrhotite separation (Finch et al., 2007a). Lead ions can sometimes be the activator of choice (Miller et al., 2006). In comparison to the Cu/sphalerite case, most metal ion activators in other systems are not held in the mineral lattice, but loosely on the surface as evidenced by their extraction using EDTA at least in the important alkaline pH range (Fuerstenau et al., 2007).

Sodium Sulfide

Oxidized minerals of lead, zinc, and copper, such as cerussite, smithsonite, azurite, and malachite, float poorly with sulfhydryl collectors due to a combination of collector loss through precipitation by heavy metal ions dissolved from the mineral lattice, and collector coatings that are weakly anchored and readily removed by particle abrasion (Fuerstenau et al., 2007). Such minerals are activated by sodium sulfide or sodium hydrosulfide by a process referred to as “sulfidization” (Fuerstenau et al., 1985; Malghan, 1986; Newell et al., 2007). Large quantities, up to 10 kg t⁻¹ of such “sulfidizers,” may be required, due to the relatively high solubilities of the oxidized minerals.

In the case of sodium sulfide, in solution it hydrolyzes and then dissociates:

${\mathrm{Na}}_2\mathrm{S}+2{\mathrm{H}}_2\mathrm{O}\leftrightarrow 2\mathrm{NaOH}+{\mathrm{H}}_2\mathrm{S}$ (12.10)

(12.10)

$\mathrm{NaOH}\leftrightarrow {\mathrm{Na}}^{+}+{\mathrm{OH}}^{-}$ (12.11)

(12.11)

${\mathrm{H}}_2\mathrm{S}\leftrightarrow {\mathrm{H}}^{+}+{\mathrm{HS}}^{-}$ (12.12)

(12.12)

${\mathrm{HS}}^{-}\leftrightarrow {\mathrm{H}}^{+}+{\mathrm{S}}^{2-}$ (12.13)

(12.13)

Since the dissociation constants of Eqs. (12.12) and (12.13) are low and that of Eq. (12.11) is high, the concentration of OH⁻ ions increases at a faster rate than that of H⁺ ions and the pulp becomes alkaline. Maintaining alkaline pH is important to avoid formation of hazardous H₂S gas. Hydrolysis and dissociation of sodium sulfide releases OH⁻, S²⁻, and HS⁻ ions into solution and these react with and modify the mineral surfaces. Sulfidization causes sulfur ions to pass into the lattice of the oxidized minerals, giving them a relatively insoluble pseudo-sulfide surface coating and allowing them to be floated by sulfhydryl collectors.

The addition of sodium sulfide must be controlled, as it is also a strong general depressant for sulfide minerals. The amount required is dependent on the pulp alkalinity, as an increase in pH causes Eqs. (12.12) and (12.13) to proceed further to the right, producing more HS⁻ and S²⁻ ions. For this reason, sodium hydrosulfide is sometimes preferred to sodium sulfide, as it does not hydrolyze and hence increase the pH. Excess sodium sulfide also removes oxygen from the pulp by the following reaction:

${\mathrm{S}}^{2-}+2{\mathrm{O}}_2={\mathrm{SO}}_4^{2-}$ (12.14)

(12.14)

Since DO is required in the pulp for the adsorption of sulfhydryl collectors (Section 12.4.4), flotation efficiency is reduced.

Sodium sulfide and hydrosulfide can be used in conjunction. In reprocessing weathered stockpiled Cu/Zn ore, Les mines Selbaie used a combination of NaHS to the SAG/ball mills (70/30 split) at pH 9 and Na₂S to the cleaner regrind at pH 5.5 to virtually restore the Cu and Zn metallurgy to that of normal ore (Finch et al., 2007a). Part of the problem was considered to be activation of sphalerite and pyrite by Cu ions released from the weathered ore. In addition to sulfidization, sulfide ions act to precipitate the Cu ions as Cu₂S, a process known as sequestration.

In processing mixed sulfide-oxidized ores, the sulfide minerals are usually floated first, followed by sulfidization of the oxidized minerals. This prevents the depression of sulfides by sodium sulfide. It has been suggested (Zhang and Poling, 1991) that the detrimental effects of residual sulfide can be eliminated by the coaddition of ammonium sulfate. Use of the relatively inexpensive ammonium sulfate appears to reduce the consumption of the more expensive sulfidizing agent while enhancing the activating effect.

Non-sulfide Systems

Because of the similarity in surface properties of many of the valuable and gangue non-sulfide minerals, they tend to respond to the same collectors and consequently the use of regulating agents is extensive (Miller et al., 2002; Nagaraj and Ravishankar, 2007). Just two examples of activation are selected as illustrations. In reverse flotation of silica from iron ores using an anionic collector, Ca ions are introduced to adsorb on the silica and alter the surface charge to positive, thus enabling collector adsorption through the electrostatic mechanism (Houot, 1983). In the flotation of feldspar from quartz with an amine (cationic) collector at acid pH, fluoride is added as activator. The fluoride selectively lowers the negative charge on the feldspar compared to the quartz, thus promoting adsorption of the cationic collector (Shergold, 1984).

Cyanide

This reagent is widely used in the selective flotation of lead–copper–zinc and copper–zinc ores to depress sphalerite, pyrite, and certain copper sulfides. Sphalerite rejection from copper concentrates is often of major concern, as zinc is a penalty element in copper smelting.

Cyanide chemistry is comprehensively reviewed by Guo et al. (2014). There are three broad categories of cyanide species: free cyanide (CN_free: HCN, CN⁻); weak acid dissociable cyanide (CN_WAD: e.g., cyanide complexes with Cu, Ni, Zn); and strong acid dissociable cyanide (CN_SAD: e.g., complexes with Fe). Guo et al. (2014) provide thermodynamic stability constants.

Accidental activation of sphalerite by Cu²⁺ ions released from copper minerals can be countered by the addition of cyanide. In this role the cyanide could be classed as a deactivator. The cyanide acts by reacting with copper in solution, first reducing Cu²⁺ to Cu⁺ then forming soluble cupro-cyanide complexes, Cu(CN)_x^1−x (e.g., Cu(CN)₂⁻), which are stable and prevent the uptake of copper by Eq. (12.9). The form of the complex depends on pH and pulp potential (Seke and Pistorius, 2006).

Cyanide needs to be added early in the circuit, for example, the grinding stage, to complex the copper ions before abstraction by sphalerite, as subsequent removal of Cu from activated sphalerite is difficult.

Cyanide is most commonly used as the sodium salt, which hydrolyzes in aqueous solution to form free alkali and relatively insoluble hydrogen cyanide:

$\mathrm{NaCN}+{\mathrm{H}}_2\mathrm{O}\leftrightarrow \mathrm{HCN}+\mathrm{NaOH}$ (12.15)

(12.15)

The hydrogen cyanide then dissociates:

$\mathrm{HCN}\leftrightarrow {\mathrm{H}}^{+}+{\mathrm{CN}}^{-}$ (12.16)

(12.16)

The dissociation constant of Eq. (12.16) is very low compared with that of Eq. (12.15), so that an increase in pulp alkalinity reduces the amount of free HCN, but increases the concentration of CN⁻ ions. An alkaline pulp is essential, as free hydrogen cyanide is extremely dangerous. The major function of the alkali, however, is to control the concentration of cyanide ions available for formation of the copper cyanide complex:

$2{\mathrm{CN}}^{-}+{\mathrm{Cu}}^{+}\leftrightarrow \mathrm{Cu}{(\mathrm{CN})}_2^{-}$ (12.17)

(12.17)

Cyanide is also used in depression of pyrite. In that case formation of a surface ferric ferrocyanide complex, which both produces a hydrophilic site and blocks a possible collector adsorption site, is the suspected mechanism (Fuerstenau et al., 2007).

Apart from the reactions of cyanide with metal ions, it can react with metal xanthates to form soluble complexes, preventing xanthate adsorption on the mineral surface. The greater the solubility of the metal xanthate in cyanide (i.e., the more soluble the metal xanthate–cyanide complex), the less stable is the attachment of the collector to the mineral. It has been shown that lead xanthates have very low solubilities in cyanide, copper xanthates are fairly soluble, while the xanthates of zinc, nickel, and iron are highly soluble. On this basis iron and zinc sulfides should be readily separated from galena in processing complex ores by the use of cyanide. In the separation of chalcopyrite from sphalerite and pyrite, close control of cyanide ion concentration is needed. Sufficient cyanide should be added only to complex the heavy metal ions in solution and to solubilize the zinc and iron xanthates. Excess cyanide forms soluble complexes with the slightly less soluble copper xanthate, depressing the chalcopyrite.

The depressive effect of cyanide depends on its concentration and on the concentration of the collector and the length of its hydrocarbon chain. The longer the chain, the greater is the stability of the metal xanthate in cyanide solutions (i.e., the soluble metal xanthate–cyanide complex is harder to form) and the higher the concentration of cyanide required to depress the mineral. Relatively low concentrations of xanthates with short hydrocarbon chains are therefore preferred for selective flotation where cyanides are used as depressants.

Cyanides, being highly toxic, must be handled with care. They also have the disadvantage of being expensive and they dissolve gold and silver. If the plant includes cyanidation to recover Au from flotation concentrates and/or tails, the use of cyanide as a depressant can be accommodated as these plants include cyanide destruction. If water from the cyanidation plant is to be recycled to the flotation section then cyanide in solution needs to be controlled. As discussed in Section 12.8, under certain conditions it is possible that cyanides can cause accidental copper activation of sphalerite. Despite these disadvantages, they are widely used. For example, Umipig et al. (2012) found that cyanide was essential in processing a Cu–Zn ore to produce separate Cu- and Zn-concentrates in the face of accidental Cu activation of the sphalerite (see also Section 12.17.3). The process includes cyanide destruction and active management of accidental activation so that CuSO₄ addition is not required.

Zinc Sulfate

While many plants function efficiently with cyanide alone, in others an additional reagent, generally zinc sulfate, is included to ensure satisfactory depression of sphalerite. The introduction of zinc ions can counter copper ions from depositing on the sphalerite surface by shifting Eq. (12.9) to the left. Accidental activation by Pb²⁺ ions is similarly countered by adding Zn²⁺ ions; if lead activation is the principal problem zinc sulfate alone may be sufficient. More complex reactions may also occur: cyanide may react with zinc sulfate to form zinc cyanide, which is relatively insoluble and precipitates on the sphalerite surface rendering it hydrophilic and blocking collector adsorption. In an alkaline pulp, zinc hydroxide forms, which is precipitated on the sphalerite surface, again rendering it hydrophilic and preventing collector adsorption.

Sulfur Dioxide

Widely used as a galena depressant in copper–lead separation, effective sphalerite rejection is also achieved in copper cleaner and copper–lead separation circuits through acidification of pulps by injection of SO₂. The addition of SO₂ is generally considered to (a) reduce the pulp potential which retards collector adsorption (Section 12.4.4) and (b) produce a series of sulfoxy species, giving a range of hydrophilic surface reaction products. In addition, SO₂ in combination with air (oxygen) can produce strongly oxidizing conditions through formation of Caro’s acid (H₂SO₅). This may result in the oxidization (destruction) of collector coatings. It does also mean that SO₂ cannot be employed when treating ores that contain secondary copper minerals, like covellite or chalcocite, since they become soluble in the presence of sulfur dioxide and the dissolved copper ions activate zinc sulfides (Konigsmann, 1985).

Sulfur dioxide does not appreciably depress chalcopyrite and other copper-bearing sulfides. In fact, adsorption of xanthate on chalcopyrite may be enhanced in the presence of SO₂, and the addition of SO₂ before xanthate results in effective sphalerite depression while increasing the floatability of chalcopyrite. The use of SO₂ in various Swedish concentrators is discussed by Broman et al. (1985), who point out that SO₂ has the advantage over cyanide in sphalerite depression in that there is little copper depression, and no dissolution of precious metals. However, it is indicated that the use of SO₂ demands adaptation of the other reagent additions, and in some cases a change of collector type is required.

In reverse flotation of pyrite from zinc concentrate a combination of heat and SO₂ has been employed, although that practice is largely abandoned today for environmental concerns. The harsh conditions this combination implies speaks to the difficulty of deactivating sphalerite once Cu has been adsorbed (Grano, 2010).

Rather than SO₂, sulfoxy species, sulfite (SO₃²⁻) and metabisulfite (HSO₃⁻), have been used. They function by a combination of formation of hydrophilic surface products and reduction in DO and pulp potential that retards collector adsorption.

Sulfide, Hydrosulfide

More than 40% of the western world’s molybdenum is produced as a by-product from porphyry copper ores. The molybdenite (MoS₂) is collected along with copper sulfides in a bulk Cu–Mo concentrate. The two minerals are then separated, almost always by depressing the copper minerals and floating the molybdenite. Sodium hydrosulfide or sodium sulfide is used most extensively, though several other inorganic compounds, such as cyanides, and Noke’s reagent (a product of the reaction of sodium hydroxide and phosphorous pentasulfide) are also used (Nagaraj et al., 1986). The action is through control of pulp potential, as discussed in Section 12.8.

Phosphates

Accidental activation by lead ions can be controlled by addition of polyphosphates, (P_nO_3n+1)⁽ⁿ⁺²⁾⁻, an inorganic complexing agent. The shorter chain phosphates function by forming a precipitate, the longer chain members by forming soluble complexes. An example of their use is the Myra Falls Cu–Pb–Zn concentrator where increasing recovery of Zn to the Cu concentrate was traced to increasing levels of galena in the ore (0.1% Pb in the original ore rising to 0.5% Pb). Addition of mono-phosphate deactivated the sphalerite (Yeomans, 2008). Provided the level of Pb ions is not too high, polyphosphates can be effective and economic (Raschi and Finch, 2006).

Natural Products

The natural products and derivatives include (Pugh, 1989; Bulatovic, 1999): starch and dextrin (polysaccharides), and carboxylmethylcellulose (CMC); tannin and quebracho; and lignosulfonates. The basic unit of starch (and dextrin) and CMC polymers is shown in Figure 12.23. Dextrins are fragmented starches of lower molecular weight (i.e., n in the Figure is smaller for dextrin compared to starch). Example uses are (Bulatovic, 2007): starch in depression of iron oxide minerals in reverse flotation of silica using amine; dextrin in depression of galena in copper–lead separation; and CMC in depression of silicate minerals in processing Ni-sulfide ores and platinum group (PGM) minerals. For starches and dextrin the functional group appears to be the OH, with adsorption a combination of H-bonding with mineral surface OH groups and possible chemical bridging via Ca²⁺ and other bivalent ions, which are known to play a role (Figure 12.24(a)) (Bulatovic, 2007). In CMC the functional group is the carboxyl, COO⁻, which may bond directly with surface metal atoms (Figure 12.24(b)) or again via Ca²⁺ bridging. Gums are another polysaccharide used largely in non-sulfide systems. Huang et al. (2012) have tested chitosan, a natural polyaminosaccharide, for chalcopyrite–galena separation.

Synthetic Products

These are largely based on polyacrylamide (PAM). The general structure of PAM-based polymers is shown in Figure 12.25. Functional groups include COO⁻, CH₂O⁻ for non-sulfides and CH₂S⁻, CS-NH₂ for sulfides. Targets for replacement are sulfur species (e.g., HS) used in Cu–Mo and Cu–Pb separation, which are objectionable reagents on safety and environmental grounds.

pH Regulators

Flotation is generally carried out in alkaline conditions, with the added advantage that most collectors, including xanthates, are stable under these conditions, and corrosion of cells, pipework, etc., is minimized. Alkalinity is controlled by the addition of lime, sodium carbonate (soda ash), and, to a lesser extent, sodium hydroxide (caustic soda) or ammonia.

Lime is by far the most common alkali used. The choice, however, depends on factors other than strictly pH. Lime by introducing Ca ions into the system is sometimes not the desirable agent. Soda ash can be favored as the carbonate CO₃²⁻ sequesters calcium and other metal ions from mineral surfaces as precipitates, which both improves particle dispersion and cleans surfaces, making the collector more effective. Espinosa (2011) describes the improved performance substituting soda ash for lime in several Peňoles Cu–Pb–Zn concentrators attributed to avoiding depression by calcium sulfate. Combined use of lime and soda ash may provide an economic compromise (Nesset et al., 1998). Sodium hydroxide is used in iron ore flotation, again to limit effects of calcium, and is employed in bitumen ore processing to aid release of surfactants that contribute to the flotation process.

To reduce pH, sulfuric or sulfurous acids (addition of SO₂) are preferred on the basis of cost. Occasionally injecting exhaust CO₂ may provide sufficient acidity. At the Niobec concentrator the cleaning section for pyrochlore recovery uses a progressive reduction in pH using hydroflurosilicic and oxalic acids (Biss, 1984).

These pH regulators are often used in significant amounts. For example, lime, although cheaper on a unit weight basis than collectors, the overall cost in sulfide mineral flotation is roughly double that of the collector used, so significant savings can be achieved by the proper choice and use of pH regulators (Fee and Klimpel, 1986).

Lime is used in the form of milk of lime, a suspension of calcium hydroxide particles in a saturated aqueous solution. Lime or soda ash is often added to the slurry prior to flotation to precipitate heavy metal ions from solution. In this sense, the alkali may be considered a deactivator, as some heavy metal ions can activate minerals such as sphalerite and pyrite. Since the heavy metal salts precipitated by the alkali can dissociate to a limited extent and thus allow ions back into solution, cyanide is often used with the alkali to complex them.

Lime can also act as a depressant for pyrite and arsenopyrite when using xanthate collectors. Both the hydroxyl and calcium ions participate in the depressive effect by the formation of mixed films of Fe(OH), FeO(OH), CaSO₄, and CaCO₃ on the surface, so reducing the adsorption of xanthate. Lime has little such effect with copper minerals, but does depress galena to some extent. In the flotation of galena, therefore, pH control may be effected by the use of soda ash, pyrite and sphalerite being depressed by cyanide.

As was shown earlier, the effectiveness of sodium cyanide and sodium sulfide is governed to such a large extent by pH that these reagents are of scarcely any value in the absence of alkalis. Where cyanide is used as a depressant, the function of the alkali is to control the cyanide ion concentration (Eqs. (12.16) and (12.17)), recognizing for each mineral and given concentration of collector there exists a “critical” cyanide ion concentration above which flotation is impossible. Laboratory results for several minerals are given in Figure 12.26. It is seen that chalcopyrite can be floated from pyrite at pH 7.5 and 30 mg L⁻¹ sodium cyanide and since, of the copper minerals, chalcopyrite lies closest to pyrite relative to the influence of alkali and cyanide, all the copper minerals will float with the chalcopyrite. Thus, by careful choice of pH and cyanide concentration, excellent separations are theoretically possible, although in practice other variables serve to make the separation more difficult. The figure shows that adsorption of xanthate by galena is uninfluenced by cyanide, the alkali alone acting as a depressant.

In addition to controlling the interaction of mineral with collector and regulating agents, pH also controls metal ion speciation in solution and in some cases collector speciation.

Metal Ion Speciation

Multivalent metals ions are hydrolysable, that is, they undergo the following series of reactions:

${\mathrm{M}}^{2+}+{\mathrm{OH}}^{-}\to \mathrm{M}{(\mathrm{OH})}^{+}+{\mathrm{OH}}^{-}\to \mathrm{M}{(\mathrm{OH})}_2(\mathrm{s})+{\mathrm{OH}}^{-}\to \mathrm{M}{(\mathrm{OH})}_3^{-}$ (12.18)

(12.18)

Species distribution diagrams can be found in a variety of sources (e.g., Miller et al., 2007). The monohydroxy complex and the hydroxide precipitate are active at the mineral surface. Their interaction is revealed by measuring zeta potential, as illustrated in Figure 12.27. Depicted for silica, the same trend holds for many minerals, non-sulfides, and unoxidized sulfides. Three charge reversal pHs are shown: the first corresponds to the IEP of the mineral (there is little interaction with the free ion); the second and third reversals reflect adsorption of the monohydroxy complex and of hydroxide precipitate, the last reversal being the IEP of the hydroxide itself. An explanation for the second charge reversal through interaction with the monohydroxy complex is illustrated in Figure 12.28 for CaOH⁻ and PbOH⁻. The resulting Ca⁺ site (Figure 12.28(a)) provides the “bridge” in Figure 12.24(a); and the Pb⁺ site (Figure 12.28(b)) is sometimes considered responsible for interaction with thiol collectors, giving accidental activation.

The uptake (adsorption) of hydroxide precipitates is an example of heterocoagulation. Over a certain pH range, the particle is negatively charged and the hydroxide precipitate (colloid) is positively charged (IEP ~ pH 10 for the “hydroxide” in Figure 12.27). With sufficient coverage, the particle surface takes on the character of the precipitate, hence the third charge reversal coincides with the hydroxide IEP. In a survey of Cu–Zn concentrators, Mg- and Zn-hydroxy species were identified as promoting particle aggregation (El-Ammouri et al., 2002). Magnesium ions are considered to be the cause of molybdenite depression when floating Cu–Mo ores in seawater at about pH 10 where the hydroxide forms (Nagaraj and Farinato, 2014b; Castro et al., 2014). Figure 12.27 shows clear evidence for possible surface contamination by metal ion species that may influence flotation.

For sulfides the trend in Figure 12.27 is seen for unoxidized surfaces; most sulfides in that state have an IEP ~ pH 2–3, similar to sulfur (Mirnezami et al., 2012). When superficially oxidized, the usual situation, the buildup of oxy-hydroxide species shifts the sulfide IEP to higher pH values. The trend in Figure 12.27 is mimicked for sulfides by progressive increases in the level of oxidation.

Interestingly, the same trends in the presence of hydrolysable cations are seen in the zeta potential of bubbles, which shows the adsorption of metal ion hydrolysis products is quite general (Han et al., 2004).

Collector Speciation

The form of collector can change with pH. An extreme example is the decomposition of xanthate at low pH. In non-sulfide mineral flotation, collectors like fatty acids and fatty amines can be in molecular (undissociated) or ionic form. The reactions are:

$\mathrm{R}–{\mathrm{COO}}^{-}+{\mathrm{H}}^{+}\leftrightarrow \mathrm{R}–\mathrm{COOH}$ (12.19)

(12.19)

$\mathrm{R}–{\mathrm{NH}}_3^{+}+{\mathrm{OH}}^{-}\leftrightarrow \mathrm{R}–{\mathrm{NH}}_2+{\mathrm{H}}_2\mathrm{O}$ (12.20)

(12.20)

In the undissociated form (to the right of the arrow), solubility is low and it is generally the ionic form that acts as a collector. There is a pH where the concentration of ionic and molecular forms is equal, the pK_diss: for fatty acids at pH < pK_diss the molecular form predominates, and for amines at pH > pK_diss the molecular form predominates. For oleate the pK_diss (or pK_a)=4.95 and for dodecylamine pK_diss (pK_b)=10.63 (Pugh, 1986). The pK_diss is thus an important parameter in the action of these collectors. For example, in Figure 12.9 (Section 12.4.3) for the amine collector, further increase in pH above ca. pH 12 results in rapid decline in flotation as the molecular form becomes dominant. Noting that the maximum flotation rate can often occur at pH near the pK_diss has suggested that combinations of ion and molecular forms, possibly ion-molecular species, can be the active collector in some instances (Pugh, 1986; Rao and Forssberg, 2007). In the fatty acid case, the complex is referred to as an acid-soap. Surface activity (e.g., measured by surface tension depression) is known to be maximum around the pK_diss (Finch and Smith, 1973; Ananthapadmanabhan and Somasundaran, 1988) thus an effect on the frother-related properties of fatty acids and amines can also be expected. Values of K_a and K_b (pK=−log K) are given by Somasundaran and Wang (2006).

Collectorless Flotation

A certain level of oxidation of sulfide minerals can result in formation of metal deficient or polysulfide sites (Figure 12.4), which can make the mineral floatable without collector. Their formation is an electrochemical process, the anodic reaction being:

$x\mathrm{MS}\leftrightarrow {\mathrm{M}}_{(x-1)}{\mathrm{S}}_x+{\mathrm{M}}^{2+}+2{\mathrm{e}}^{-}$ (12.21)

(12.21)

with the common electron acceptor again being oxygen. The reaction can continue to eventually form elemental sulfur (S₈), which is strongly hydrophobic. Referred to as self-induced or collectorless flotation it has attracted attention (Luttrell and Yoon, 1984; Chander, 1991; Ralston, 1991). Although oxygen is the usual electron acceptor there are others which could substitute, notably iron (Fe³⁺/Fe²⁺) ubiquitous in sulfide flotation pulps. In copper activation of sphalerite it is noted that some level of collectorless flotation is achieved, which is attributed to Cu²⁺ acting as an additional electron acceptor to oxygen and intensifying the oxidation (Lascelles and Finch, 2002). The addition of sulfide ions can also bring about collectorless flotation (Yoon, 1981). Pyrrhotite is in effect a metal deficient iron sulfide (Pratt et al., 1994) and naturally has polysulfides on the surface and often exhibits collectorless flotation.

Collectorless flotation can be demonstrated in the plant practice, but there is usually little incentive to exploit as there is no easy way to determine if collectorless flotation will be robust to ore changes and collector addition removes that concern (Leroux et al., 1994).

In the disintegration of fibers involving strong acid attack mentioned in Section 12.6.6 with respect to a novel processing route for ultramafic Ni ores, the oxidizing conditions produced by the concentrated sulfuric acid rendered the pentlandite (and pyrrhotite) collectorlessly floatable. At the same time the amount of MgSO₄ put into solution provided an ionic strength sufficient to replace the need for frother (see Section 12.5.4). The process, therefore, gave a technically interesting collectorless-and-frotherless flotation system.

Galvanic Interaction

Recalling from Section 12.4.4 that sulfides have different rest potential values means that when they are in contact with each other in electrolyte (i.e., process water) it will induce galvanic interaction. Iwasaki and co-workers (Adam and Iwasaki, 1984; Nakazawa and Iwasaki, 1985) were among the first to consider the implications of galvanic interactions in flotation.

To illustrate galvanic interaction consider three materials in a grinding mill, in order of decreasing rest potential: pyrite, galena, and mild steel. The associated current–potential diagram is shown in Figure 12.30. In this case, each anodic reaction is matched with its own cathodic (oxygen reduction) reaction. The mixed potential at I_A=I_C establishes that the highest anodic current is contributed by oxidation of the steel, and that the highest cathodic current is contributed by reduction of oxygen on the surface of pyrite, with galena oxidation/reduction reactions in between. The physical picture is illustrated in Figure 12.31: the pyrite draws electrons from both steel and galena accelerating their oxidation.

The result of the galvanic interactions depicted in Figure 12.31 includes lowering of DO (consumed by accepting electrons) and release of metal ions both with implications for flotation. The lowering of DO in the grinding mill discharge may retard reaction with collector and reduce floatability. Aeration in a bank of flotation cells may restore the DO level necessary for flotation but there may be a case for preflotation aeration. In reactor–separator flotation cells (Section 12.13.3) with short retention time, ensuring that the DO level is sufficient for flotation is an important precaution.

Oxidation promoted by galvanic interaction releases Pb and Fe ions into the pulp. (Recent work has shown that galvanic interactions, especially in the presence of pyrite, promote formation of hydrogen peroxide (H₂O₂), which further accelerates oxidation rates (Nooshabadi et al., 2013).) The released Pb²⁺ ions may accidentally activate other minerals. Acting on the suspicion that Pb²⁺ ions were activating pyrite, Brunswick Mine installed a flotation stage between the autogenous (AG) mill and the ball mill to remove galena before the further size reduction (i.e., increase in galena surface area) added to the release of more Pb²⁺ ions. The result was increased pyrite rejection from the bulk Cu–Pb concentrate (Finch et al., 2007a). Addition of reducing agents into the grinding circuit to reduce galena oxidation was another successful option at another plant (Finch et al., 2007a).

Oxidation of the steel (i.e., corrosion) releases a series of hydrophilic iron oxy-hydroxide species into the pulp that adsorb indiscriminately and act as general depressants. Trying to limit contamination of flotation pulps by iron species has led to the use of inert (non-steel) media in stirred regrind mills (Pease et al., 2006), and is an argument in favor of AG milling (Bruckard et al., 2011). One reason raising pH is common practice is to retard release of iron species from the steel through suppression of galvanic interactions by inhibiting the reduction of oxygen on the pyrite (in the example in Figure 12.31). Replacing forged steel media by corrosion-resistant high chrome media at Ernest Henry Mining permitted the pH to be lowered in the roughers from 10.6 to 9.6 without release of excessive iron corrosion species (as determined by low Fe extraction by EDTA). The copper and gold metallurgy was essentially unchanged with the reduction in lime consumption valued at $700,000 (Kirkwood et al., 2014). Of course this needs to more than offset any increase in grinding media costs. The conversion was the result of extensive laboratory testing using the Magotteau grinding mill which permits close control over pulp chemistry. Details on laboratory testing to simulate plant chemical conditions are given in Grano (2010).

The importance of simulating the plant chemical environment when conducting laboratory tests is illustrated by investigations into cases of apparent accidental activation of sphalerite by cyanide in Cu–Zn ores (Rao et al., 2011). At low pulp potential, copper leached into solution is held as a cupro-cyanide complex (Section 12.6.3), which is not only stable but with Cu in the +1 state means that there is no exchange with Zn²⁺ ions in the sphalerite surface lattice. If the pulp potential is raised, however, the cupro-cyanide is oxidized to cupric cyanide, a reaction that both increases leaching of copper and puts Cu in the +2 state, making exchange with Zn²⁺ ions in the sphalerite lattice now possible. The implication is that if laboratory grinding is with steel media giving low pulp potential then this accidental activation mechanism is not observed; but if the plant grinding circuit is AG then pulp potential is higher than in the lab experience and accidental activation of sphalerite may occur. Espinosa (2011) reviewing plant practices at several Cu–Zn operations reported sphalerite activation by cyanide. In tests using the Magotteau mill to control pulp chemistry, he established the role of oxidation in the use of cyanide causing accidental activation of sphalerite. Alternative depressants such as sulfoxy species may have to be substituted if this cyanide activation problem is encountered.

The impact of galvanic interactions can be seen in ore stockpiles and reclaimed pillars at the end of mine life. Pyrite in weathered (i.e., oxidized) ore will still be relatively untarnished and bright, while other sulfides like sphalerite may disintegrate to the touch.

Mild steel usually has the lowest rest potential and is oxidized by most sulfide minerals. There are exceptions. In milling matte at Vale Base Metals’ Sudbury operation the steel balls are not corroded. The sulfide minerals in the matte are chalcocite (Cu₂S) and heazelwoodite (Ni₃S₂) and these have a lower rest potential than the steel (Bozkurt et al., 1994); in this case the steel is “protected” and accelerates oxidation of the minerals.

Galvanic interactions are not necessarily detrimental as the preceding may imply. Considering that xanthate adsorption is an oxidation reaction, the adsorption of xanthate on galena in contact with pyrite is promoted, while at the same time the surface of pyrite hosts the formation of hydroxyl ions which generate hydrophilic (depressant) Fe oxy-hydroxide products. The galvanic contact thus promotes selective flotation of galena over pyrite. In the Galvanox™ process pyrite is deliberately added to a second sulfide to enhance leaching rates, an example where galvanic effects are exploited to advantage. There are options to include the Galvanox™ process in combination with flotation to form an integrated circuit (Dixon et al., 2007).

Use of Nitrogen

There is occasional reference to the potential for gases other than air in sulfide mineral flotation, nitrogen being the logical candidate based on cost (Johnson, 1988). The action of nitrogen substituting for air as flotation gas can be interpreted as reducing galvanic interactions. It has been shown that nitrogen promotes xanthate flotation of pyrite relative to other sulfide minerals (Martin et al., 1989; Xu et al., 1995). The postulated mechanism is that the nitrogen breaks the galvanic interaction between sphalerite and pyrite, which suppresses the formation of oxy-hydroxides on the pyrite surface. The removal of these hydrophilic species both eliminates a pyrite depressant and opens sites for adsorption of collector.

A commercial process for pyrite reverse flotation from sphalerite concentrate employs heat and SO₂ conditioning. The combination appears to first produce strong oxidizing conditions that destroy the Cu–xanthate on the sphalerite surface and ties up the Cu as an oxide, and finally leads to a low pulp potential that favors pyrite flotation. The use of nitrogen helps maintain the low potential, thus permitting less aggressive conditioning (Xu et al. 1995; Grano, 2010). In pyrite reverse flotation from Pb concentrate in Brunswick Mine’s Pb upgrading circuit, heating alone (to ca. 75°) appears to be sufficient (Martin et al., 2000); this could in part be due to heating lowering the DO sufficient to suppress galvanic contact and thus prevent the oxy-hydroxide from forming on the pyrite surface.

Experiments on pyrite alone show that pyrite is depressed by nitrogen, explained by the rest potential being too low to support formation of dixanthogen (Figure 12.18). When pyrite is in combination with other sulfides this is not the case. The nature of the hydrophobic species on the pyrite at low pulp potential remains the subject of investigation (Miller et al., 2006). The situation with the other major iron sulfide mineral, pyrrhotite, is different. Using nitrogen pyrrhotite is effectively depressed in separation from pentlandite. In this case it appears that the collectorless flotation of pyrrhotite due to oxidation of the surface polysulfides to elemental sulfur is suppressed by the lowered potential with nitrogen (Finch et al., 1996, 2007a).

Practical applications of nitrogen must meet the challenge of maintaining a nitrogen atmosphere in face of the many opportunities for air (oxygen) to enter the system, via open balls mills and hydrocyclones being two examples. This ingress of oxygen has hampered commercial attempts in the promising use of nitrogen to depress pyrrhotite away from pentlandite. One possible solution is to enclose flotation cells and recycle the air which leads to an atmosphere enriched in nitrogen by taking advantage of the high oxygen demand of sulfide pulps. A small supply of nitrogen would still be needed to replace the volume of oxygen consumed to maintain ambient air pressure inside the cell. Another possibility is to take advantage of locations where a nitrogen stream is available, for example, off-gas from oxygen production. The patented N₂TEC technology implemented at Lone Tree mine used such an off-gas source of nitrogen and reported improved selective recovery of Au-bearing pyrite (Simmons, 1997). Where nitrogen is commonly used is in chalcopyrite–molybdenite separation, where it both reduces side oxidation reactions that consume the chalcopyrite depressants, NaHS/Na₂S, and helps maintain a low pulp potential, which depresses chalcopyrite (Aravena, 1987; Poorkani and Banisi, 2005).

To take advantage of pulp potential in control of mineral separations requires measurement. Measurement of E_p is more problematic than measurement of pH. Traditionally, it is measured using a noble metal (usually platinum but sometimes gold) as the sensing electrode with calomel (Hg/Hg₂Cl₂) or silver/silver chloride (Ag/AgCl) as reference electrodes. The sensing electrode needs to be inert and designed to include contributions to the pulp potential from the particles—it is the potential at the mineral/solution interface that matters in flotation. Sensing electrodes are prone to fouling and need close attention (Chander, 2003). An example of successful E_p measurement is in control of NaSH/Na₂S additions in chalcopyrite–molybdenite separation. Here the noble metal sensing electrode becomes coated with sulfur and responds to the sulfide ions in solution (Woods, 2010).

Control of potential through reagents is often not easy, the systems being buffered to change. The possibility of direct control through design of an electrochemical cell, successfully used at the laboratory scale, is an area of research for industrial application (Panayotov and Panayotova, 2013).