A Collector Promoter for Apatite Flotation in the Serra do Salitre Complex

Abstract

The concern about enhancing the productivity of Salitre phosphate mines has led to an extensive research and development program on new reagents, aimed at the sustainable processing of the ore, with greater production of phosphate concentrate and, consequently, less waste disposal in the tailings dam. Flotation is the most widely applied technique for processing of phosphate rocks, using mainly fatty acids and their salts as collectors. In partnership with BASF Mining Solutions, a new collector system was developed based on soy oil fatty acid (SOFA) and BASF’s co-collector Lupromin^® FP A 1210 Base (L.1210, a mix of synthetic surfactants). The laboratory scale flotation results showed that most effective performance was achieved with a ratio SOFA:L.1210 of 70:30, while in the industrial application the ideal proportion was 90:10. This difference in reagent ratio is directly related to the surface tension of the bulk, making the apatite more or less hydrophobic, directly affecting the flotation performance. These proportions were tested by comparing different reagent conditions and pH, while contact angle and surface tension studies provided a theoretical framework to indicate a more effective collector adsorption, allowing for further process optimization. Applying 10% of L.1210 on an industrial scale, a more stable industrial process was observed to the oscillations of the processing plant, corroborating that by using the collector promoter there is a significant performance increase in this circuit of the Serra do Salitre Complex, which increased, on an industrial scale, at least 7.1% in P₂O₅ recovery and reduced 12.5% of the collector dosage compared to the use of only fatty acid. In this study, it was possible to identify two main process conditions to further optimize the current reagent system applied in industrial scale: pH 9.5 with the mixture SOFA/L.1210 90/10 (θ = 75.8°); and pH 7.5 with SOFA/L.1210 70/30 (θ = 76.8°).

1. Introduction

In order to achieve the specifications required by the fertilizer industry, low-grade phosphate ores are commonly subjected to mineral processing, highlighting that the processing route depends on the geological origin (sedimentary or igneous) and the nature of the gangue minerals that will be segregated from apatite—the main phosphate-bearing mineral [1]. In case of igneous phosphate ores, direct anionic flotation is the most applied method to promote the separation between silicate minerals and apatite, and the collectors traditionally used are fatty acids, which are classified as surfactants [2]. Of these, the most widely used are rich in oleic acid, or sodium oleate soap, which comprises a monocarboxylic acid with a long C18 hydrocarbon chain and a double bond; and/or rich in linoleic acid, as well as its sodium linoleate soap, which has a C18 hydrocarbon chain with two double bonds, indicating better performance in certain applications [3]. For flotation processes, constituent species are transferred from solid to solution until electrochemical equilibrium is reached. In addition, semi-soluble minerals, such as calcite, apatite, and fluorite, when dissolved in water, undergo several hydrolysis and complex formation reactions, releasing a variety of ions in solution [4]. The non-stoichiometric dissolution of apatite increases rapidly for pH values lower than 8, due to the increased hydrolysis of surface ions. At pH lower than 7, the measurement of the zeta potential of apatite becomes complex due to the higher solubility of the mineral. The surface has an excess of positive charge, which is most likely acquired by the adsorption of Ca²⁺ ions [5].

There is a distinction between three mechanisms of interaction between collector sodium oleate and calcium-bearing semi-soluble minerals that involve specific adsorption. Such mechanisms are chemisorption, surface reaction (physical adsorption), and precipitation in the bulk (precipitation of calcium oleate within the solution). Considering the similar surface properties, it is hard to separate these minerals, which tend to float together. Many parameters influence selective flotation and adsorption mechanisms, and they can be evaluated in terms of wettability (contact angle) and surface tension characteristics [6]. According to Sis and Chander, knowledge of the wettability characteristics of minerals helps to understand the separation efficiency of the mineral of interest from the gangue minerals. Wettability involves the interaction between a liquid in contact with a solid and is directly related to its surface tension. When a drop of a liquid spreads on a solid surface, the contact angle is small, and the surface is classified as hydrophilic. When the drop shrinks over the solid, the contact angle increases, and the surface has a hydrophobic characteristic [7].

Surface tension is a thermodynamic component directly related to adhesion efficiency, relating a balance of forces between the existing interfaces in flotation. It is noteworthy that its magnitude decreases with increasing a surfactant concentration, until it reaches a minimum level, beyond which the increase in concentration does not promote a reduction in surface tension, since energetically the surfactant molecules are distributed in the medium of the solution, being available for the formation of micelles. This point of minimal surface tension at which micelles are formed is also known as critical micellar concentration (CMC) [8]. In this context, another relevant point to be highlighted is the critical concentration of wettability (CCW), which would be the minimum surface tension (or critical tension, γc) in which the wettability of a mineral surface is reversed, since tensions below γc mean that the surface energy becomes so low that, even with an adsorbed collector, the mineral may have a hydrophilic character [9]. The critical concentration of wettability can be defined by the intercept of the horizontal line cos θ = 1 with the extrapolated straight-line plot cos θ vs. γLV°, and it was denoted by γC [10].

Despite the traditional application of vegetable collectors in phosphate flotation, the selectivity promoted by fatty acids, especially in the presence of slimes, is insufficient to achieve economically viable recovery levels [11]. This situation is aggravated when using recirculated water with high amounts of Ca²⁺ ions in the pulp, which consumes the fatty acid and precipitate calcium oleate in the solution, generating a total increase in collector consumption. In the current work, the results, and characteristics of a new tailor-made collector, Lupromin^® FP A 1210 Base (a mix of synthetic surfactants, BASF property), are presented for the Salitre ore. During development, bench-scale flotation tests were carried out, where different proportions of vegetable/synthetic collector were compared and adjustments were made in the formula, finally creating a mixture that promoted a noticeable improvement in industrial flotation performance. This work aims to explain the action of collector mixtures used in two main concentrations: in the bench scale (10 mg/L) and in the industrial plant (160 mg/L), as well as to explain the reasons behind changes in component proportion differences between laboratory and industrial scale.

2. Materials and Methods

2.1. Mining Process and Mineralogy Characterization

The apatite of the Salitre complex occurs in the form of free and well crystallized grains, or in mixed particles associated mainly with silicates (pyroxene, grenade, and clay-minerals). The ore is extracted in an open pit mine through mechanical rock breaking. After that, the ore feeds a primary crushing and then stacking and reclaiming. The beneficiation plant presents a secondary crushing, primary grinding, classification through hydrocyclones, secondary grinding, desliming, conditioning, three flotation stages (rougher, cleaner, and recleaner) where only recirculated water is used, and finally magnetic separation and filtration. Both grinding stages are done in an open circuit.

The ore sample used in the bench flotation tests is a blend of phosphate ore typologies from the “Salitre I” deposit, which can be seen in Figure 1. The chemical composition determined by X-ray fluorescence (XRF) indicates that the ore is composed of a 4.51% P₂O₅ and CaO/P₂O₅ ratio of 3.24, as shown in

Table 1. Chemical composition (XRF, % mass) sample.

P2O5	CaO	SiO2	Al2O3	Fe2O3	MgO	TiO2	BaO	PF	CaO/P2O5
4.51	14.6	32.1	4.43	19.3	7.21	8.62	<0.10	5.28	3.24

. The loss on ignition of 5.28% was determined gravimetrically. The mineralogical study using the MLA—Mineral Liberation Analyzer (MLA-FEI) software (LCT-USP, São Paulo, Brazil), coupled to the scanning electron microscope Quanta 650 FEG and EDS Esprit (Bruker), indicates a mass composition of 8% apatite, 31% hydrobiotite, and 18% pyroxene, in addition to other gangue minerals, as stated in

Table 2. Mineral composition (total sample: −0.50 + 0.020 mm) through MEV/EDS-MLA.

Mineral	%	Mineral	%
Hydrobiotite	31.0	Granate	6.1
Pyroxene	18.0	Perovskite	3.4
Apatite	8.1	Other oxides	3.0
Magnetite	8.1	K-feldspar	2.8
Titanite	7.5	Quartz	1.5
Clay minerals 12–15 A	6.6	Others	4.8

and represented by Figure 2.

In Figure 3 it is possible to observe a particle that was chemically analyzed by MEV-EDS. It is a pyroxene-apatite mixed particle. Cations Ca²⁺ and Mg²⁺ in alcaline pH add positive surface charges to apatite and gangue minerals (quartz and pyroxene) [12]. It means that the presence of pyroxenes as diopside in the process requires the addition of different surfactants to permit a suitable flotation performance, otherwise, the dosage of fatty acid could be substantial, and the process would show poor selectivity.

Figure 3. Fraction −0.50 + 0.30 mm in the MEV image. The regions assigned with numbers were chemically analyzed by EDS and their composition in terms of F, MgO, SiO₂, P₂O₅, CaO, TiO₂, and Fe₂O₃ is shown in Table 3. Mixed particle pyroxene (1,2) and apatite (3).

Figure 3. Fraction −0.50 + 0.30 mm in the MEV image. The regions assigned with numbers were chemically analyzed by EDS and their composition in terms of F, MgO, SiO₂, P₂O₅, CaO, TiO₂, and Fe₂O₃ is shown in Table 3. Mixed particle pyroxene (1,2) and apatite (3).

Table 3. Chemical composition of the regions assigned in the MEV image shown in Figure 3.

EDS Region	Element/Mineral Grade (wt%)
	F	MgO	SiO2	P2O5	CaO	TiO2	Fe2O3
1	-	11.4	51.4	-	22.6	0.70	11.4
2	-	11.3	51.1	-	21.7	0.51	12.9
3	2.85	-	-	40.5	53.3	-	-

Table 3. Chemical composition of the regions assigned in the MEV image shown in Figure 3.

EDS Region	Element/Mineral Grade (wt%)
	F	MgO	SiO2	P2O5	CaO	TiO2	Fe2O3
1	-	11.4	51.4	-	22.6	0.70	11.4
2	-	11.3	51.1	-	21.7	0.51	12.9
3	2.85	-	-	40.5	53.3	-	-

2.2. Size Distribution of the Sample

The ore sample was subjected to wet milling in a laboratory jar mill, with a solids concentration of 60% for a period of 5 min. The granulometric analysis of the grinded ore was performed in wet vibrating sieve where the retained fractions were dried in an oven at a temperature of 70 °C and weighed on a precision balance. The flotation feed reached a P80 of 150 µm, and chemical analysis by fractions through XRF determined that approximately 73% of the P₂O₅ contained was distributed between the fractions −0.147 and +0.025 mm, as shown in Figure 4 [13,14].

2.3. Soy Oil Fatty Acid (SOFA) Characterization

Soybean fatty acid applied as a collector was characterized by gas chromatography, using Agilent equipment, model 7890N. The results of this characterization are shown in

Table 4. Fatty acid collector characterization.

Density at 20 °C (g/cm3)	0.91
Density at 50 °C (g/cm3)	0.89
Acidic index (mgKOH/g)	154.11
Iodine index (gI2/100 g)	112.75
Saponification index (mgKOH/g)	196.13
Free fatty acid (%)	78.58
Chain distribution (%)
C6	0.18
C8	0.20
C10	0.14
C12	0.16
C14	0.38
C16	16.48
C18	5.38
C18:1	28.64
C18:2	44.17
C18:3	4.26
Total	100.00

.

2.4. Contact Angle Measurements

The contact angle was measured on polycrystalline and polished apatite surfaces, to obtain information on their wettability by the following liquids and mixtures: fatty acid emulsion, fatty acid emulsion, and Lupromin^® FP A 1210 Base, in the proportions of 90:10 and 70:30, respectively. The tests were carried out on sample surfaces of pure apatite mineral cut along perpendicular planes to guarantee the verification of the influence of preferential crystalline orientations. The wetting angles were observed, photographed, and measured through the QuantDim software, as shown in Figure 5. The measurement conditions of the working environment were 22 °C ± 1 °C, and the average relative humidity was 55% ± 3%.

2.5. Reagents Surface Tension Analysis

The Du Noüy ring method was used to measure the surface tension of collector solutions and emulsions. The platinum ring was first cleaned with acetone and then with distilled water. Before each measurement, the surface tension of the distilled water was measured as a control. In each test, 40 mL of sample was used. The pH was maintained at the desired value with NaOH and HCl solution at 1% (w/w). The surface tension analysis of the reagents was performed using a Lauda tensiometer, at 25 °C, in triplicate.

2.6. A Qualitative Trend of Collector Adsorption through Surface Tension

A mass of 250 g of the flotation feed was introduced into collector solutions at the desired concentrations at pH 8.5. The procedure used was the same as for bench flotation: 5 min of conditioning with gelatinized cassava starch and another 5 min with the collector. After 10 min of conditioning, the suspension was filtered, and the liquid was used for surface tension analysis.

The surface tension measurements of the collector solutions were also tested to obtain the trends of surface tension reduction, mainly caused by the adsorption of the collector onto apatite mineral.

2.7. Bench Scale Column Flotation Tests

A depressant based on cassava starch gelatinized with NaOH and diluted to 3% in water was used for this work. SOFA was used, varying the proportion at 100%, 90%, and 70%, while the proportion of Lupromin^® FP A 1210 was 0%, 10%, and 30%, respectively. After SOFA saponification (5 g) carried out with 7.15 g of 10% NaOH solution (m/m), the synthetic collector was incorporated, and the mixture was diluted to 1%. The conditioning of the ore sample was carried out in a mechanical cell for a period of 5 min for each reagent, at a rotation of 900 rpm. A dry mass of 500 g feed at4 5% solids percentage was used. The depressant was dosed at 450 g/t, and the collector was added at different dosages, depending on the reagent used. The pH was adjusted to 8.5 by the addition of 5% NaOH (w/w) solution.

The flotation tests were performed in a bench column with dimensions of 5000 mm in height and 50 mm in diameter, in a single stage. The air and wash water flow conditions were, respectively, 100 nL/h and 0.15 L/min. The air was introduced through a porous surface installed at the bottom entrance of the column and pressurized air was injected upstream.

The products obtained in the flotation were filtered and dried in an oven at a temperature of 70 °C. Recovery was calculated using the dry mass of concentrate and tailings and their respective P₂O₅ contents.

3. Results and Discussion

3.1. Comparison between SOFA and Collectors’ Mixture Composition

In this section, SOFA flotation performance and the mixtures with Lupromin^® FP A 1210 Base were evaluated through lab column flotation tests, seeking a concentrate with a grade around 32.5% P₂O₅. The results were represented in Figure 6. P₂O₅ grade and recovery were plotted in the graph, as well as the standard deviation for each condition, which was carried out in triplicate.

The SOFA test presented an average grade of 32.6% ± 0.96% in the concentrate, and recovery of 63.8% ± 1.16%, demanding a dosage of 10.2 mg/L (160 g/t) of collector. The test with the mixture containing 10% L.1210 reached 33.8 ± 0.06% P₂O₅ with a recovery of 68.3 ± 0.86%, applying a collector dosage of 8.9 mg/L (140 g/t), which represents a metallurgical recovery increase of 7.1% and collector reduction of 12.5%. The test with the mixture containing 30% L.1210 reached a P₂O₅ grade of 33.3% ± 1.21% and recovery of 69.4% ± 1.14%, keeping the collector dosage at 8.9 mg/L, which means an increase of 8.8% in the recovery, when compared with the test with 100% SOFA. The flotation times were for SOFA: 1′26″, 90% SOFA + 10% L.1210: 1′39″ and 70% SOFA + 30%L.1210: 1′40″.

3.2. pH Effect on the Surface Tension and Contact Angle

Surface tension and contact angle analysis were conducted to understand the pH influence in reagents adsorption on the interface liquid/air, as well as the solid–liquid phase. The surface tension was measured as a function of collector concentration, in natural pH (7.5), considering four scenarios: SOFA, Lupromin^® FP A 1210 Base, and two mixtures between both, one with 90% SOFA and another with 70% SOFA. It is verified in Figure 7 that L.1210 promotes a reduction in surface tension, which is significantly higher than the one reached with only fatty acid. This is because the fatty acid is partially neutralized (75%), and the remaining molecule forms an emulsion in the presence of oleate ions, sodium linoleate, among other hydrocarbon chains, raising the material surface tension. L.1210 reaches CMC around 50 mg/L and then the curve is kept constant in a surface tension value around 27 mN/m. In the fatty acid curve, the CMC is not well defined, and the collector dosage needed is 10 times higher to diminish the surface tension observed in the first. Therefore, it can be affirmed that a lower CMC of the collector occurs in the presence of the L.1210 base.

In sequence, surface tension variation was measured as a function of pH, as shown in Figure 8. In Figure 8a, the collector concentration applied in the laboratory was t: 10 mg/L. It is possible to verify that L.1210 itself provides a lower and more stable surface tension (varying from 36.7 to 39.3 mN/m) than observed in fatty acid (44.7 to 58.5 mN/m), over the whole pH range, with values around 39 mN/m. The SOFA presents a significatively higher surface tension with a higher pH starting at 8.5, reaching 58.5 mN/m at pH 11.5. In the mixtures, the surface tension at pH 7.5, 9.5, and 10.5 reach intermediate values compared to those reached with the two separated products.

In Figure 8b, the surface tension was measured in the industrial process at a concentration of 160 mg/L where traditionally a higher collector dosage is demanded in relation to the bench scale. In this condition, as expected, lower values of surface tension are observed in all scenarios. Pure L.1210 shows constant values over the whole pH range studied, with values around 27 mN/m. The SOFA suffers almost no value variation in the surface tension between pH 8.5 (34.4 mN/m) and 9.5 (34.3 mN/m).

To evaluate the apatite surface susceptibility to applied collectors, the contact angle formed between sample surface and the interface bubble-particle was measured as a function of pH for the mentioned proportions of the collector concentration: SOFA, mixture 90% SOFA + 10% L. 1210 and mixture 70% SOFA + 30% L. 1210. In Figure 9a, it is possible to observe the contact angles measurements for bench scale with a concentration of 10 mg/L. The contact angles produced by SOFA are, for all pH values (except pH 11.5), lower than those obtained in the mixtures with L. 1210. For the alkaline pH range, the mixture 70/30 (SOFA/L.1210) presented the highest contact angles (pH 7.5 = 68.5° to pH 9.5 = 76.5°), with a general tendency of higher hydrophobicity with increasing pH.

In Figure 9b, the same measurements mentioned above were performed, although in the collector concentration of the industrial plant, at 160 mg/L. A high amplitude occurs among the contact angle values for the mixture 70/30 (SOFA/L.1210), indicating strong sensitivity to pH variation. In the mixture 90/10 (SOFA/L.1210), a tendency of the contact angle increasing is observed when the pH is increased until 9.5 (75.8°), which means that the particle turns more hydrophobic in this direction, showing to be the most promising scenario. For SOFA, the pH that promotes the highest contact angle is 11.5 (59.5°), a condition already used in Salitre in the past, but considering the high NaOH consumption, it was not economically viable.

Besides the contact angle measurements (θ) on the apatite surface, surface tension measurements were performed after the collector’s adsorption in the apatite to verify the tendency with pH variation in the different dosage conditions; the results are presented in Figure 10.

In Figure 10a,b, it is possible to compare the obtained results with SOFA. In Figure 10a, with a collector dosage of 10 mg/L, in pH 8.5 (44.9 mN/m), it is possible to observe the lowering of the surface tension when compared to the obtained values at pH 7.5 (53.4 mN/m) and at pH 9.5 (54.1 mN/m). For contact angle values, this reduction is not observed with the same tendency, but it occurs as expressive diminution in the contact angle at pH 9.5 (θ = 35.5°), while at pH 7.5 (θ = 47.7°) and 8.5 (θ = 47.0°), the angles remain stable. It is known that fatty acids form vesicles at a pH between 8 and 9 [15], emphasizing that these vesicles are multilamellar and their bilayers are stabilized by ionization of a proportion of the fatty acid groups through weak linkages -OH-O- in the vesicle surface [16]. At a pH around 7.5, there is a phase change of the fatty acid drops into vesicles, while at a pH around 9.5, another phase happens, changing the vesicles to micelles; the fatty acids eliminate the electrostatic repulsions that the anion polar heads of fatty acid experience in micelles, and both structures are present also in the bilayers of the vesicle [17]. In Figure 10a, the collector dosage is lower when compared with the industrial reality, the fatty acid collector is not selective enough to increase the apatite hydrophobicity, and the contact angles are lower in comparison to Figure 10b. Figure 10b shows a higher collector concentration/dosage, as the two curves have the same tendency, and it is understood that a pH of 7.5 and 9.5 is more efficient to concentrate the apatite, because it allows for a higher quantity of the collector to be adsorbed on its surface, which can be confirmed through the higher surface tension and contact angle.

Comparing Figure 10a with Figure 10c, applying 10% of co-collector Lupromin^® FP A 1210 Base, it is possible to analyze the following: a flattening in the surface tension curve, tension diminution at pH 7.5 (53.4 → 50.1 mN/m) and 9.5 (54.1 → 53.1 mN/m), and a higher tension at pH 8.5 (44.9 → 46.6 mN/m). It happens because the surfactant effect in L. 1210 is rupturing the vesicles that can be formed at pH 8.5, increasing the collector effect, which can be verified through the higher contact angle value (47.0° → 56.5°). Comparing Figure 10b,d, the flattening in the surface tension curve can also be seen, although an expressive increment in the contact angle value in the three studied pH points is observed. These results demonstrate that for the mixture 90:10, the best pH condition to be practiced in industrial scale would be pH = 9.5.

In Figure 10e, applying the proportion 70:30 in a bench scale dosage, it is still possible to observe a performance drop at pH 8.5, supported by a lower contact angle (θ = 63.0°) in relation to the other pH (7.5 com 68.5°; 9.5 com 76.5°), but with values above that observed both with the pure collector (35.5° ≤ θ ≤ 47.7°) and in the 90:10 ratio (40.5° ≤ θ ≤ 56.5°). In Figure 10f, with the increasing concentration of the collector, the highest contact angle value found is at pH 7.5 (76.8°). At pH 8.5, there is again an impact that impairs apatite flotation, as the contact angle decreases with a drop in surface tension, indicating a reduction in collector adsorption on the apatite surface, as more surfactant leaves the mineral surface and returns to the bulk of the solution. It is possible that the result with the highest floatability potential is at pH 7.5, as it is the analyzed pH that contains the greatest amount of fatty acid in molecular format, requiring the co-collector to facilitate its emulsification. Therefore, it is understood that depending on the pH of the process and the proportion of surfactants in the mixture, the process could be optimized if carried out at a pH close to natural in the proportion 70:30. The results showed that a higher proportion of L.1210 favors mass recovery, increased production, and recovery despite the small reduction in the quality of the final concentrate.

From this study, according to industrial results, the use of L.1210 with at least 10% proportion is essential for achieving the recovery target of 60% P₂O₅, ensuring a higher level of production and recovery mass with an acceptable quality of the final concentrate, in which the value of 32.47% P₂O₅ grade obtained meets the current specification.

4. Conclusions

In this work, the behavior between fatty acid-based collector and the surfactant Lupromin^® FP A 1210 Base was evaluated in the apatite flotation of the Serra do Salitre Complex. Through this study and the experiments carried out, an increase in the metallurgical performance in the recovery of P₂O₅ was observed, highlighting that flotation tests carried out in the laboratory indicate that the optimum performance comes from a SOFA:L.1210 ratio of 70:30 (an 8.8% increase in P₂O₅ recovery), while in the experiments carried out on an industrial scale, the ideal proportion tested so far was 90:10, denoting an increase in P₂O₅ recovery of approximately 7.1%.

Extensive discussions were presented, comparing performance trends of pure fatty acid and mixtures with L.1210 at different ratios and different pH, complementing this information with measurements of contact angle and surface tension. The effect of pH was determined, showing that the modified collector with ratios from 90:10 to 70:30 (SOFA:L.1210) allows flotation to be conducted at 7.5 ≤ pH ≤ 9.5 with optimized performance, bringing operational flexibility for production engineers at the beneficiation plant. Operating in a natural pH condition can reduce the process complexity and increase the operational safety. Further industrial tests will be carried out in the future to validate these alternative conditions, aiming to further increase the performance of the apatite flotation at Serra do Salitre: (a) pH 9.5 with the mixture SOFA/L.1210 90/10 showed a contact angle of 75.8°; (b) pH 7.5 with the mixture SOFA/L.1210 70/30 showed a contact angle of 76.8°.

The work carried out here showed that the processing of phosphate-bearing minerals can be improved through combinations of anionic collectors with suitable special surfactants such as L.1210. The convergence of optimization between the chemical and mining industries favored the performance of flotation, which is underscored by the increase of at least 7.1% in the metallurgical recovery of P₂O₅ along with a 12.5% reduction in reagent consumption. There are possibilities of improvement after identifying the process conditions that may optimize the performance in Serra do Salitre on an industrial scale.