Review of the Main Factors Affecting the Flotation of Phosphate Ores

: The way to successfully upgrade a phosphate ore is based on the full understanding of its mineralogy, minerals surface properties, minerals distribution and liberation. The conception of a treatment process consists of choosing the proper operations with an adequate succession depending on the ore properties. Usually, froth flotation takes place in phosphate enrichment processes, since it is cheap, convenient, and well developed. Nevertheless, it is a complex technique as it depends on the mineral’s superficial properties in aqueous solutions. Aspects such as wettability, surface charge, zeta potential, and the solubility of minerals play a basic role in defining the flotation conditions. These aspects range from the reagents type and dosage to the pH of the pulp. Other variables namely particles size, froth stability, and bubbles size play critical roles during the treatment, as well. The overall aim is to control the selectivity and recovery of the process. The following review is an attempt to add to previous works gathering phosphate froth flotation data. In that sense, the relevant parameters of phosphate ores flotation are discussed while focusing on apatite, calcite, dolomite, and quartz as main constituent minerals.


Introduction
Phosphate ores are the only pronounced source of phosphorus. As a crucial component with high economic importance, phosphorus remains indispensable in the phosphoric acid industry, fertilizers, and elemental phosphorus production. The demand on phosphate has increased throughout the years. A total of 47 million tons P2O5 were consumed worldwide in 2019. This consumption is predicted to increase up to 50 million tons in 2023 [1]. As presented in Table 1, the worldwide leaders of phosphate mine production are China (110 million tons/y), Morocco (36 million tons), United States (23 million tons/y), and Russia (14 million tons/y). Countries such as Jordan, Saudi Arabia, Vietnam, Brazil, Egypt, Australia, and Senegal produce between 1 and 8 million tons/y. Phosphate mine production is expected to grow, especially in Jordan, Morocco, Saudi Arabia, and Senegal following projects expansions [1].
Conventionally, phosphate deposits are classified into five major categories: Marine sedimentary, igneous, metamorphic, weathering sedimentary, and biogenic [2]. Sedimentary deposits are predominant, providing 80% of the world's phosphate production. These deposit's grades range typically between 10% and 30% P2O5 and can be upgraded to (30-35%) P2O5. Morocco, to renovate the particles surfaces. The remaining silica is then floated using an amine reagent [12]. For calcareous ores, separation by flotation of carbonates (calcite and dolomite) and phosphates is challenging since the two mineral groups share similar surface chemical properties. There are also colloidal phosphate ores, sedimentary ores containing, in general, a very small grain sized apatite (cryptocrystal). The beneficiation of these ores is challenging. It requires fine grinding and the use of selective reagents since apatite is incorporated with gangue minerals, forming "collophanite".
With the depletion of rich phosphate deposits of simple compositions and the abundance of complex poor phosphate reserves, research in phosphate froth flotation keeps evolving. At this stage, efforts are directed toward understanding what affects phosphate flotation and the different aspects that result in the success or the failure of a flotation operation. With these regards, the aim of this review is to focus on the factors affecting the flotation of sedimentary phosphate ores, as they provide 80% of the world's phosphate production. Table 1. Leading phosphate producers [13].   Up to 3% [7,14] Low Al2O3 content improves the filtration rate by promoting the growth of gypsum crystals [15]. Reduces corrosion caused by fluoride ion [14].
High Fe2O3 content causes excessive sludge formation, decreases the filtration rate, and influences the acid viscosity [14].
High SiO2 content causes wear and erosion of equipment and can impair the filtration of gypsum [7,15].

Cadmium
Carbonate apatite (Cd substitutes Ca and/or is trapped in the structure during its formation by the sedimentation of phosphate rock) [16].
Does not pose notable problems in the production of phosphoric acid [7]. Toxic in specific end products (animal feed and fertilizers) [15].

Uranium
Following the sedimentation of the phosphate ore, U might substitute Ca in the apatite crystal and/or is adsorbed into it or forms uranium phosphate minerals such as phosphuranylite [18].
Can be recovered as a by-product [14].

Phosphate and Gangue Minerals' Surface Properties
Mineral surface properties are critical for flotation control as they represent the fundamental concepts behind the technological process. They allow to, both, predict and describe the physical and chemical superficial interactions of the mineral-reagent, mineral-medium, and mineral-mineral interactions.

Phosphate Minerals' Solubility and Surface Charge
The surface charge of phosphate minerals and their impurities (i.e., gangue minerals) is affected by many variables, e.g., pH and RPO values (reduction/oxidation potential). Respectively, the mechanisms governing the surface charge are complex rendering separation by froth flotation challenging in some cases. Minerals from the apatite group are the most abundant. It has been suggested that their active surface sites are formed by calcium hydroxyl (≡CaOH) and phosphorus hydroxyl (≡POH) groups [19]. It is considered that ≡Ca─OH 2+ and ≡P─O − are dominant surface groups [19]. The ions H + and OH − are specifically adsorbed on the surface. Basically, depending on the pH, two different surface groups (both metal ions and ligands) can be adsorbed on the hydroxylated apatite surface. Hence, the adsorption of the flotation modifying agents on mineral surfaces is usually expressed in relation to pH being a critical parameter. The apatite's surface charge depends not only on pH but also on the concentration of various chemical species as well as on pretreatments. In this context, hydrogen, hydroxyl, calcium, fluoride, and phosphate are the apatite's potential determining ions [20]:  H + and OH − ions are critical as they control the solution pH and eventually the mineral's surface charge [19];  Phosphate (PO4 3-) decreases the mineral's surface charge under all pH conditions [20];  Calcium makes the mineral more positively charged under all pH conditions [20];  Fluoride is found to increase the surface charge in acidic medium and slightly decreases it in basic solutions following the possible formation of fluorite (CaF2) and fluorapatite (Ca5(PO4)3F), respectively [20].
The zeta potential expresses the electrical potential difference between the ion layer adsorbed onto the particle's surface and the bulk solution [21]. It enables prediction as to whether or not a reagent will bind to the particle's surface. The isoelectric point (IEP) is the pH value corresponding to a zero zeta potential. IEP values of hydroxyapatite, fluorapatite, francolite, and collophane in different operating conditions are listed in Table 4. Parameters such as mineral purity, particles size, solid content, electrolyte, and measurement technique influence the IEP value.

Gangue Minerals' Solubility and Surface Charge
The most common impurities found in the sedimentary phosphate ores are carbonates (calcite and dolomite) and quartz.


Calcite solubility and surface charge Calcite solubility in aqueous medium leads to the release or deposition of Ca 2+ and CO3 2− on mineral surfaces depending on the solution pH. This latter and concentrations of Ca 2+ , CO3 2− , and HCO3 might fluctuate under the influence of dissolved atmospheric CO2. Moreover, the hydrated calcite surface can have both protonated anion sites (>CO3H) and hydroxylated cation sites (>CaOH). Hence, in a froth flotation system, the solution pH controls the calcite surface charge. Nevertheless, it has been demonstrated experimentally that, when the calcium ions concentration was kept constant, the calcite surface charge was independent of pH [22]. The adsorption of Ca 2+ and CO3 2− onto the mineral surface ( Figure 1) is represented through the following surface complexation Reactions (with or without the presence of CO2) [22].
To summarize, calcite surface determining ions (PDI) are the lattice ions Ca 2+ and CO3 2− . Mg 2+ can also be considered as a PDI. The pH affects the zeta potential by moderating the equilibrium pCa for a given CO2 partial pressure (pCO2), therefore, it does not directly control the zeta potential [22]. Table 5 presents calcite IEP values and their measurement operating conditions.    Dolomite solubility and surface charge The surface of dolomite resembles that of calcite. Dolomite is soluble in aqueous medium and the lattice ions Ca 2+ , Mg 2+ , and CO3 2− are susceptible, depending on the solution pH, to either dissolve in the solution or precipitate on mineral surfaces. Moreover, the solution pH and concentrations of Ca 2+ , Mg 2+ , CO3 2− , and HCO3 might fluctuate under the influence of dissolved atmospheric CO2. The species existing on the dolomite surface were investigated by some researchers. Figure 2 presents the calculated speciation at the dolomite-solution interface [44]. Based on these calculations, at a pH below 4, carbonate sites are protonated with predominantly >CO3H species. In higher pH conditions, >CO 3− dominates as deprotonation occurs. pH as well as the dissolved carbonate concentration influence the speciation at the dolomite metal sites. At a pH below 8, >MeOH 2+ species are dominant, however, once the pH exceeds 8 > MeCO 3− dominates. The study confirms the similarities between calcite and dolomite interfaces in aqueous solutions. It also indicates that the PDI for the dolomite surface are its lattice ions (as for calcite) being Ca 2+ , Mg 2+ , and CO3 2− [44]. Table 6 presents IEP values of dolomite and their measurement conditions.   Quartz solubility and surface charge Quartz can be found as an impurity in most phosphate ores. Its superficial properties in aqueous solution resembles the properties of amorphous silica. Therefore, the two materials carry similar functional surface groups, and consequently, have the same reaction mechanism [46]. In the quartzfatty acid flotation system, and in the absence of metal ions, the predicted Reactions are: >SiOH  SiO − + H + , (4) and RCOOH  RCOO − + H + . (5) According to the reactions above, the adsorption of fatty acids on the quartz surface is nearly impossible. Though, in the presence of metal ions, quartz hydrophobicity can be achieved by surface reactions such as the ones presented below (Reaction 6, 7, and 8) [46]. Table 7 presents the IEP values of quartz and its measurement operating conditions. >SiOH + Me 2+  SiOMe + + H + , >SiOH + Me 2+ + RCOO − + H2O  SiOMeOHRCOO − + 2H + , Mineral wettability is critical to the establishment of an efficient froth flotation process. It conveys the mineral's floatability with or without the addition of collectors. The wettability is expressed by an empirical angle. It can be experimentally measured by various methods. For instance, a liquid drop is laid on a solid surface. The intersection of the liquid-solid and liquid-vapor interfaces forms the angle termed "contact angle". A hydrophobic surface retains a contact angle larger than 90°. A hydrophilic surface has a smaller contact angle [48]. The contact angle was first introduced by Tomas Young (1805) and can be calculated through its equation: γlv cos θY = γsv − γsl, (9) where γlv, γsv, and γsl represent the liquid-vapor, solid-vapor, and solid-liquid interfacial tensions, respectively, and θY is the Young's contact angle.
However, the contact angles calculated through young's equation are usually different from the measured ones. The calculated θY can be experimentally confirmed only in ideal conditions where the solid surface is physically and chemically homogenous, and the experiment is conducted under extremely controlled conditions. In reality, most measurements are done on heterogenous imperfect solid surfaces (especially for minerals). The advancing and the receding contact angles are other variations of θ. They are measured by expanding and contracting the liquid, respectively and the difference between these two is called the hysteresis (H), a very interesting and informative parameter: Table 8 presents phosphate minerals, calcite, dolomite, and quartz contact angles with water using different experimental methods. As already mentioned, the contact angle's precision depends on the sample preparation as its interface needs to be smooth and homogeneous and the measurement must be conducted in controlled conditions to avoid vibrations.

Flotation Reagents
Flotation reagents, especially regulators (activators, depressants, and pH regulators) and collectors, impact the selective separation of phosphate minerals from impurities, which are usually carbonates (calcite and dolomite) and silicates (quartz). During the phosphate direct flotation process, gangue minerals are depressed and the phosphate mineral retaining a hydrophobic surface is floated. The opposite occurs during reverse flotation. Table 9 summarizes the reagents used in the phosphate reverse and direct flotation.


Calcite and Dolomite Reagents Generally, anionic collectors are used for carbonates flotation. In the past decades and still in some cases today, anionic collectors' production depended on tall oil and oxidized petroleum as raw materials. However, these come with multiple downsides. Rather, vegetable oils such as rice bran, hydrogenated soybean, cottonseed, and jojoba oils represent a promising inexpensive source for fatty acids [59]. There are some cases when amphoteric collectors (e.g., aminopripionic acid [60,61], carboxyethyl imidazoline [62]) are used for floating carbonates. Moreover, nonionic collectors are generally used to improve the performance of ionic surfactants. They reduce the electrostatic repulsions between ionic head groups, generate hydrophobic chain interactions, and consequently, the adsorption of the ionic collector on mineral surfaces [29]. Table 10 presents the conditions and results where anionic and amphoteric collectors are used for the flotation of calcite and dolomite.
According to research conducted on different organic reagents used as dolomite depressants, carboxymethyl cellulose, citric acid, and naphtyl anthyl sulfonates are effective [63]. The β-naphthyl sulfonate formaldehyde condensate (NSFC) was used as a dolomite depressant during the anionic flotation of collophane at pH value of 9 [34]. NSFC was chemically adsorbed on the dolomite surface and barely adsorbed to collophane. However, NSFC's chemical toxicity is a severe limitation [34]. Additionally, Bacillus subtilis and Mycobacterium phlei have been tested as depressants of apatite and dolomite [64]. Results indicated the adsorption of both bacteria species on the minerals surfaces. They functioned as depressants for dolomite, as well as for apatite. Although these bacteria do not ensure apatite dolomite selectivity, a simulation of the flotation environment can provide a better choice of bacteria for this purpose.

 Quartz Reagents
Usually, the elimination of silicates minerals, such as quartz, is conducted using cationic collectors. Table 11 presents different reagents mentioned as silica collectors in the literature. Sodium silicate is reported as a performing quartz depressant. Silva et al. [65] investigated the mechanism of quartz depression using sodium silicate. They have found that at a pH value of 7, monomeric Si(OH)4 and polymeric species, resulting of sodium silicate hydrolysis, were adsorbed on the quartz surface. For a sample containing 97.2% SiO2, the quartz depression was observed at a dosage of 1000 g/t Na₂SiO₃. Optimum floatability was obtained in the pH range from 5 to 8 using a dosage of 1500 g/t Na₂SiO₃ in the presence of 150 g/t amine. Under these conditions, quartz floatability was less than 10%. At a pH value of 11, floatability is high, suggesting that sodium silicate is not adsorbed onto the quartz surface. Morever, starch exhibits a depressive effect on quartz flotation. Its flocculation property enables it to adsorb onto the quartz surface engendering its depression [66]. Nevertheless, starch is not an efficient depressant for quartz due to its low selectivity. It is still used in a number of concentrators as it is cheap in comparison to other effective reagents [67]. IRR: Impurities-removal-ratio.

Apatite Minerals Reagents
Apatite minerals are usually depressed in acidic medium. Studies showed that at a pH value below 4.5, apatite floatability is weak due to the dominance of Ca 2+ , CaH2PO4 + , and H2PO4 − . The H2PO4 − species in the suspension occupies apatite's active sites and leads to a poor recovery. Therefore, inorganic acids are considered effective depressants with phosphoric acid being the most used as it does not cause any complication during the enrichment process. Phosphate salts can depress apatite, as well.
Always in acidic medium, using sulfate or oxalate salts is reported to further depress apatite. These reagents incite the precipitation of the dissolved Ca 2+ , leading to apatite dissolution and a greater presence of phosphate ions in the suspension [75,76].
Additionally, a synthetic polymeric depressant "ACCO-PHOS 950" was developed by Cytec. The aim is to limit apatite loss during silicates flotation using amine collectors. The reagent proved to be effective when used on North Africa's high-grade phosphate ores [77]. Zhang and Snow (2014) [78] investigated sodium tripolyphosphate, fluosilicic acid, diphosphonic acid, starch and sodium silicate as apatite depressants candidates. The phosphate ore used contains 19.82% P2O5 and 40.28% Insol. Sodium tripolyphosphate gave promising results. Using it as apatite depressant yielded a final phosphate concentrate assaying approximately 31% P2O5 and up to 7% Insol with over 94% P2O5 recovery. During reverse flotation of silica from a fine-grained feed, starch effectively depressed the apatite.

Influence of Particle Size
Particle size is a main parameter in the flotation process. Particles of various sizes behave differently in the flotation system, directly affecting the recovery, selectivity, and overall performance. Usually, fine particles' flotation is less preferred compared to coarse mineral particles. Fine-grained minerals have a higher surface energy, leading to a nonselective reagent consumption, entrainment or entrapment of particles and an instability of bubble-particle aggregates. A study of the relationship between the phosphate particles size and flotation recovery was carried out by Gaudin et al. (1931). The results revealed that under similar operating conditions, particles of various sizes exhibit different flotation kinetics. The maximum phosphate minerals recovery was obtained when treating size fractions between 60 and 200 µm. Mineral liberation, which is related to the particle size, is also a parameter in the froth flotation of phosphate ores. Fine grinding is critical to attain greater liberation but will unfavorably affect the separation efficiency. Therefore, there is generally a compromise when choosing the particle size. A preforming flotation system requires an effective liberation of desired minerals without overgrinding the ore.

Froth Stability
The attachment of hydrophobic mineral particles to gas bubbles leads to the genesis of a concentrated mineral lather. The froth's stability is critical and is generally maintained using foaming agents (frothers). The froth must sustain the particle's weight, while resisting excessive coalescence and bursting. On the other hand, the proper functioning of the flotation process requires a manageable and easily suppressed froth for a practical recovery [79]. Froth stability is affected by the parameters such as gas flow rate, stirring rate, solid content, particle size, mineralogy, reagents type and dosage, and water's ionic strength. Farrokhpay et al. [80] noted that froth stability can be assessed through parameters such as froth half-life time, froth maximum height at equilibrium, dynamic froth stability factor, bubble growth across the froth phase, air recovery, and solid loading on bubbles on top of the froth surface, froth velocity, and froth rise velocity, etc.
Particle size and solid content: Liu et al. [79] investigated, in their recent study, the effect of particle size and solid content on fluorapatite, calcite, dolomite, and quartz froths stability. Using a setup inspired from Lunkenheimer's report [81], they assessed the dynamic froth stability factor (Ʃ) for different particle sizes and solid contents. This factor is calculated using: Vf is the froth volume (cm 3 ), Q is the gas flowrate (cm 3 /s), S is the cross-sectional area of the column (cm 2 ), and Hmax is the maximum froth equilibrium height (cm). Liu et al. [79] found that for fluorapatite, calcite, dolomite, and quartz the fine particles contribute to froth stability as a result of capillary mechanisms, whilst larger particles lead to froth rapture and instability. Additionally, they observed that increasing the solid content to a certain extent stabilizes the froth (Figure 3). Water's ionic strength/mineralogy: According to Liu et al.'s study [79], dolomite and especially the calcite froths stability were found to be superior than those of fluorapatite and quartz. It was demonstrated that Ca 2+ and Mg 2+ ions resulting from calcite and dolomite dissolution stabilize the froth. Moreover, the froth half-life time increased dramatically with the addition of Ca 2+ ions up to 5.10 −4 mol/ L Ca 2+ but decreased rapidly thereafter. Reagent type: The reverse flotation process widely used for upgrading siliceous phosphate ores are considered challenging, especially regarding froth stability. In most cases, cationic collectors are used to float the siliceous gangue. However, they produce an expanding overly stable froth that might entrain other minerals, causing lack of selectivity [59,79,82].

Bubble's Size
Hoang et al. [83] studied the effect of flotation time and froth height on the bubble's size. They found that the bubble's size increased with flotation time, contributing to the reagent's consumption and the decrease in the solid content leading to bubble coarsening. They also stated that the bubble's size increased with the froth height due to water film rupture resulting from low reagents concentration and hydrophobic particles existing in the water film.

Conclusions
Based on common scientific data, it can be deduced that the phosphate ores' enrichment was not as complicated as the enrichment of other minerals, namely, metal sulfides and oxides. The deposits already exploited have been high in P2O5 content, contained less impurities, and there were not as many strict quality regulations as there is today. With the depletion of these deposits, upgrading complex phosphate ores, while respecting rigorous criteria, is more and more challenging. Phosphate processing regulations are not only about the concentrate's quality. They also cover environmental issues, such as CO2 emission, polluting and toxic reagents, water and waste management.

•
Huge carbon dioxide emissions are the main downside of thermal beneficiation techniques such as calcination. Hence, froth flotation was developed to replace it in many cases. • Water recycling is already applied in most phosphate concentrator plants. • Phosphate waste valorization is an interesting recent topic with numerous industrial opportunities. Hakkou et al. [84] mentioned multiple methods of phosphate wastes valorization. One way is the use of alkaline phosphate wastes (APW) to inhibit the acid mine drainage (AMD). Regarding its high calcite content, 15% APW was used to neutralize the acidity produced by pyrrhotite tailings' oxidation [85]. The APW were also assessed in the passive AMD water treatment [84]. Additionally, phosphate wastes with a size less than 1 mm, were tested in storeand-release (SR) covers to reclaim industrial mine sites [86]. • Flotation reagents can be environmentally harmful and highly toxic. For instance, different apatite depressants (organic and inorganic) are used in the reverse flotation of sedimentary phosphate ores. The most common ones are phosphoric and sulfuric acid and their derivatives. Using these inorganic depressants can entail, however, potential threats to the environment [87] (e.g., calcium phosphate scale formation and water eutrophication [88]). Organic depressants have been developed for apatite/carbonate separation, as well. Nevertheless, they are usually extremely toxic which limits their use [87]. Funding: This research was financially supported through the research program between OCP S.A under the specific agreement OCP/UM6P AS34.

Conflicts of Interest:
The authors declare no conflict of interest.