Dry Concentration of Phosphate Ore by Using a Triboelectrostatic Belt Separator in Pilot Scale

Abstract

This study investigates the feasibility of using a triboelectrostatic belt separator (TBS) as a dry alternative to conventional magnetic separation for concentrating apatite from a phyllosilicate-rich phosphate ore from the Unidade de Mineração de Angico, Brazil. The testing material contained 22.9% P₂O₅ and exhibited over 90% mineral liberation even at coarse fractions (+0.6 mm), being mainly composed of apatite and Mg/Al-bearing phyllosilicates. Pilot-scale experiments were carried out in an M6c TBS, evaluating operational parameters such as electrode polarity, belt speed, feed rate, and electrode gap. In the rougher stage, apatite’s positive charging behavior enabled separation from negatively charged gangue, with optimal conditions (run 4) producing a concentrate of 25.3% P₂O₅ at 85.1% recovery. Cleaner experiments further upgraded product quality, with runs 15 and 18 yielding concentrates of 29.0% and 28.9% P₂O₅ and overall P₂O₅ recoveries of 69.3% and 74.5%, respectively. Compared to high-intensity magnetic separation currently applied at the industrial plant, the TBS achieved superior mass and P₂O₅ recoveries and more effective MgO removal, although Fe₂O₃ and Al₂O₃ contents remained slightly above market thresholds. These results confirm the technical feasibility of triboelectrostatic separation for phosphate beneficiation, offering environmental benefits through reduced water consumption and tailings generation. Further research should focus on finer particle sizes (−0.3 mm), electrode design, and surface charge modifiers to enhance industrial performance.

1. Background

Apatite is a phosphorus-bearing mineral that typically occurs in phosphate ores in association with several gangue phases, including carbonates (e.g., calcite, dolomite), silicates (e.g., quartz, mica, pyroxene, amphibole), and oxides (e.g., magnetite, hematite, ilmenite, anatase) [1,2,3]. Direct utilization of raw phosphate ores in fertilizer production is uncommon, as beneficiation is generally required. Typical processing routes involve comminution, classification, and concentration techniques such as magnetic separation (MS) and froth flotation [3,4,5]. For more than eight decades, low-grade ores have been upgraded worldwide by froth flotation, which requires chemical reagents (surfactants, polymers, pH modifiers) and generates a substantial quantity of wet tailings, demanding environmentally responsible disposal [5,6]. Increasing restrictions on water availability in recent decades [7] have intensified the demand for dry concentration alternatives to conventional wet methods. Within this context, triboelectrostatic separation (TES) has been investigated for over 90 years, primarily at laboratory scale, as a potential dry route for separating apatite from silicate and/or carbonate gangue phases [8].

This study investigates the dry concentration of a phosphate ore from the Unidade de Mineração de Angico (UMA), operated by Fosnor (Fertilizantes Fosfatados do Norte-Nordeste LTDA), located in Angico dos Dias, Bahia, Brazil [9,10]. Due to the semiarid conditions of the region, phosphate beneficiation at UMA has, for over two decades, relied on a three-stage MS circuit: low-intensity MS (~0.05 T) for magnetite removal, high-intensity MS (~0.5 T) for hematite and ilmenite removal, and very high-intensity MS (>1.0 T) for the rejection of Fe²⁺-bearing phyllosilicates (Figure 1a). As mining advanced to deeper sections of the open pit, newly exploited ores enriched in non-magnetic phyllosilicates began to be fed into the plant, reducing the quality of the apatite concentrate. In these cases, R₂O₃ (Al₂O₃ + Fe₂O₃) contents exceeded 5%, while P₂O₅ grades remained within the acceptable range of 28%–30%. To address this limitation, TES was considered as an alternative to replace the two high-intensity (>0.05 T) MS stages currently used in the UMA beneficiation plant (Figure 1b). The objective of this work is to evaluate the feasibility of applying TES to this phosphate ore and to analyze the effects of operating parameters of a pilot-scale triboelectrostatic belt separator (TBS) on producing apatite concentrates with P₂O₅ contents above 28.5% while maximizing recovery. These findings contribute to both practical and theoretical advances on the separation of apatite–phyllosilicates, bridging the gap between laboratory research and industrial-scale implementation, while addressing innovative mineral processing technologies and offering an environmentally favorable solution for ore beneficiation.

1.1. Triboelectrostatic Separation

Triboelectrostatic separation (TES) exploits differences in the triboelectric charging behavior of non-conductive materials to achieve mineral separation according to the magnitude and polarity of the acquired surface charge [11]. Like froth flotation (

Table 1. Froth flotation versus triboelectrostatic separation (TES).

Attributes	Froth Flotation [10]	TES [11,12,13,14]
Separating medium	Water	Air
Differentiating property	Wettability by water	Fermi level/work function
Variable that controls the differentiating property	Contact angle	Density and sign of the acquired surface charge.
Surface modification previous to separation	Conditioning with chemical reagents (collectors, frothers, modifiers)	Contact/friction between mineral/mineral, mineral/polymers, mineral/walls of equipment, assisted or not by gas adsorption and radiation.
Modus operandi of the mineral separation	Hydrophobic particles collide and adhere to air bubbles and float; Hydrophilic particles do not adhere to air bubbles and sink	Negatively charged particles move to the positively charged electrode, vice versa.

), TES can be regarded as a surface-property-based process [11,12,13,14,15,16], in contrast to other beneficiation technologies that rely on bulk properties, such as magnetic susceptibility in magnetic separation or density in gravity separation [17]. Because surface properties are amenable to modification, the effectiveness of both froth flotation [12] and TES [11,14,15,16] strongly depends on appropriate surface conditioning to enhance selectivity. Nevertheless, as highlighted by Mirkowska et al. [11], the fundamental triboelectrification mechanisms that control TES selectivity remain insufficiently understood, making the prediction of separation performance and the optimization of process parameters particularly challenging. Within this context, the present study seeks to advance the understanding of the operational variables of a triboelectrostatic belt separator (TBS).

The selectivity of triboelectrostatic separation (TES) arises from differences in the ability of solid materials to acquire and retain static surface charges. The charge transferred between two materials upon contact can be predicted by their respective work functions (W_F), defined as the minimum energy required to remove an electron from a solid surface [13,14,18,19,20]. As illustrated in Figure 2, when two particles come into contact or are rubbed against each other, the particle with the higher W_F becomes negatively charged (blue particle), while the particle with the lower WF acquires a positive charge (orange particle). This electron transfer (triboelectrification) occurs through contact and frictional charging mechanisms, which may result from collision, sliding, or vibration at particle interfaces [18]. Because triboelectrification is strongly influenced by the presence of adsorbed water at the solid–gas interface, and since adsorbed water equilibrates with ambient humidity, it is essential to maintain minimal air moisture during TES operations [21,22].

The relative tendency of a mineral species to acquire positive or negative charge can be estimated using the “triboelectric series,” an empirically derived ranking in which materials with lower W_F (more likely to acquire positive charge) are positioned at the top, while those with higher W_F (more likely to acquire negative charge) are positioned at the bottom [23]. In laboratory-scale studies using an electrostatic plate separator, Ferguson [24], following the earlier approach of Fraas [25], determined the relative positions of several minerals within such a triboelectric series (

Table 2. Examples of triboelectric series developed for non-conductive minerals, based on the intensity of the positive (+) or negative (−) acquired charge.

Intensity and Relative Signs of the Acquired Charge	Triboelectrostatic Series
	Fraas [25]	Ferguson [24]
++++++++++++++	Siderite
++++++++++++	Olivine
+++++++++++	Andracite
++++++++++	Apatite	Apatite
+++++++++	Nepheline	Carbonates
++++++++	Magnesite	Monazite
+++++++	Allanite	Titanomagnetite
+++++	Staurolite	Ilmenite
++++	Beryl	Rutile
+++	Grossularite	Leucoxene
++	Eudialyte	Magnetite/Hematite
+	Sphene	Spinel
−	Stilbite	Garnet
−−	Netafite	Staurolite
−−−	Diopside	Altered Ilmenite
−−−−	Cryolite	Goethite
−−−−−−	Hornblende	Zircon
−−−−−−−	Monazite	Epidote
−−−−−−−−	Chromite (Spinel)	Tremolite
−−−−−−−−−	Euxenite	Hydrous Silicates
−−−−−−−−−−−	Scheelite	Aluminosilicates
−−−−−−−−−−−−−	Microcline	Tourmaline
−−−−−−−−−−−−−−	Albite	Actinolite
−−−−−−−−−−−−−−−	Quartz	Pyroxene
−−−−−−−−−−−−−−−−−	Rinodolite	Titanite
−−−−−−−−−−−−−−−−−−−	Actinolite	Feldspar
−−−−−−−−−−−−−−−−−−−−−	Hexagonite	Quartz
−−−−−−−−−−−−−−−−−−−−−−−	Glauconite

). The intensity of the acquired charge is represented by the numbers of positive or negative signs shown in Table 2. In these rankings, apatite consistently appears at the top, reflecting its strong tendency to donate electrons (lower W_F) upon contact with common gangue minerals such as quartz and hydrous/aluminum silicates (e.g., mica, clays), which have higher W_F.

Once mineral particles are triboelectrically charged, they respond to an external electric field established between two parallel electrodes of opposite polarity. Separation is achieved because positively and negatively charged particles follow distinct trajectories determined by the magnitude and sign of their acquired surface charges. As summarized in

Table 3. Forces acting on a particle that settles in triboelectrostatic separators and their contribution towards the selectivity of the apatite/silicates separation.

Forces	Cause	Effect	Contribution to the Selectivity
Electrostatic	Magnitude, density, and sign of mineral surface charge prior to separation [26]; Magnitude and polarity of the static field generated by the electrodes positioned within the separators [26,27].	Attractive/repulsive forces between mineral particles and the equipment’s electrodes [26,27].	Different position of apatite versus silicates in the triboelectric series maintained by either Ferguson [24] or Fraas [25].
Gravity	Volume (size) and specific gravity ($\rho$) of the particles involved in the separation [26,27].	Larger, heavier particles settle more quickly than smaller, lighter ones despite their buoyancy [26,27].	Apatite particles (p > 3000 kg/m3) are denser than many silicates (p < 2800 kg/m3) [28].
Drag	Particles’ shape and air temperature [26,27].	Particles of lower sphericity factor ($\mathsf{\Psi }$) settle at a lower rate than rounded particles [26,27,29].	Apatite particles settle as rough tetrahedrons (0.62 < Ψ < 0.64), whereas mica and clay particles show a platy-shaped habit (Ψ < 0.3) [29].

, particle motion in the separator is governed by the combined action of electrostatic forces, gravitational forces, and aerodynamic drag exerted by the dry air medium. The interplay of these three forces enables selective separation between apatite and associated gangue minerals [26,27].

Two main types of triboelectrostatic separators are currently available on the market for industrial operation: free-fall separators (FFSs) [30] and triboelectrostatic belt separators (TBSs) [31]. As illustrated in Figure 3, FFS devices operate with vertically positioned electrodes, whereas TBSs employ horizontal electrodes. This configuration in the TBS allows the operator to control whether charged particles migrate upward or downward within the equipment by adjusting the polarity of the upper and lower electrodes.

A further distinction between the two systems lies in the mechanism of charge acquisition. In the FFS, particle charging must be induced prior to separation, typically through external devices such as pneumatic cyclones, fluidized beds, or vibrating chutes and cones [32]. In contrast, the TBS simultaneously induces and separates charges, since frictional contact occurs both between particles and between the particle surface and the moving belt surface [22], promoting the transference of electrons. Consequently, charge generation in the TBS begins at the feed distributor, where the particles starts to contact each other, and continues throughout the separation process.

The TBS has been employed industrially to remove unburned carbon from fly ash in coal-fired power plants for the production of concrete-grade pozzolans used as a cement substitute [33], as well as in the processing of industrial minerals [34,35]. Unlike the FFS, which is generally restricted to particle sizes greater than 75 µm, the TBS is effective across a much wider particle size range, from very fine (<1 µm) to moderately coarse (up to 500 µm) [31].

1.2. Factors Influencing the Separation Process

The performance of triboelectrostatic separation can be optimized through the appropriate configuration of operational parameters in the TBS, including feed port position, electrode polarity, feed rate, electrode gap, and belt speed, as detailed in Section 3.2 for the model employed in this study.

Beyond these parameters, several physical and operational factors significantly affect separation efficiency. Mirkowska et al. [11] compiled a comprehensive list of variables influencing TES, including the physicochemical properties of particle surfaces (composition, dopants, roughness, adsorption of molecules or ions), charge transfer mechanisms (via electrons, ions, or mass exchange), and material history effects (grinding, aging, storage). Environmental conditions, particularly temperature, humidity, and gas atmosphere, are also known to play a decisive role [11]. In this study, particular attention was given to the following factors:

• Temperature of the processed material: The charging behavior of phosphate minerals is temperature-dependent, which directly affects their separation from silicate and carbonate gangue minerals [36].
• Electric field geometry: Asymmetries in electrode configuration may produce low-intensity regions, insufficient to deflect denser particles, thereby reducing selectivity [11].
• Turbulence in the inter-electrode gap: Aerodynamic instabilities between electrodes can alter particle trajectories, leading to drag-related losses and cross-contamination of product streams [11].
• Residence time: Prolonged residence times may increase particle accumulation on electrodes, causing overloading, while turbulence induced by the counter-current belt motion may dislodge weakly adhered particles. Both phenomena reduce recovery efficiency and separation selectivity [11].
• Residual particle moisture: Surface moisture interferes with triboelectric charging mechanisms, diminishing their effectiveness [11].
• Particle size and shape distribution: Morphology and granulometry affect the mass-to-surface ratio, thereby influencing particle responses to both aerodynamic drag and electrostatic forces [11].
• Charging via belt contact: Heavier particles exert greater pressure on the belt surface and on adjacent particles, enhancing triboelectric charging through intensified contact. The belt’s material properties and surface conditions strongly affect particle polarization and, consequently, separation efficiency [11].

In TBS systems with horizontally aligned parallel electrodes, such as the model used in this study, electrode polarity configuration exerts a major influence on performance. Alternating the polarity of the upper and lower electrodes can substantially modify particle trajectories, particularly when density contrasts exist within the mixture. For instance, if target particles are denser and attracted to the upper electrode, their trajectories are governed by the interplay between gravitational forces, which drive them downward, and electrostatic forces, which pull them upward. Additional aerodynamic drag induced by belt motion, as well as fluctuations in feed point and particle movement within the separation zone, can further destabilize particle paths and compromise efficiency. Conversely, when denser particles are directed toward the lower electrode, gravitational and electrostatic forces act in the same direction, generally producing more stable and predictable separations.

The opening in the upper electrode, designed to accommodate the feed inlets, may introduce asymmetry in the electric field, potentially affecting electrode performance. [11] Monitoring the electric current in the electrodes provides valuable insights into the dynamic behavior of particles within the separator, particularly with respect to trajectory randomness and separation stability. Such observations can be interpreted using the Shockley–Ramo theorem, which relates the induced current on an electrode to the motion of charged particles under an applied electric field [37,38], as expressed in Equation (1).

$$\mathrm{i}=\mathrm{q}·\mathrm{V}·{\mathrm{E}}_{\mathrm{w}}$$

(1)

where

$q=$ charge of the particle;

$V=$ velocity vector of the particle;

$E_w=$ weighting field (a function of the geometry and position of the electrodes).

While particles travel within the electric field, their velocity directly contributes to the induced current measured at the electrodes. Erratic or unstable trajectories manifest as random fluctuations in this current, reflecting instability in the separation process. Such behavior indicates a loss of control over dynamic operating conditions and compromises the selectivity of triboelectrostatic separation.

Beyond trajectory irregularities, a further factor that reduces the current measured at the electrodes is the accumulation of insulating particles on their surfaces. Mineral mixtures frequently contain high-resistivity phases, which can alter the local electric field distribution in the vicinity of the electrodes. This phenomenon is comparable to that observed in electrostatic precipitators, where non-conductive dust forms a dielectric coating over the electrode surface, leading to a decline in collected current [39].

In this scenario, adhered particles act as a dielectric layer at the electrode–free space interface, distorting the local electric field and suppressing the induced current. Such a layer creates a capacitive barrier that not only reduces the interaction between newly charged particles and the electrodes but also diminishes the effective electric field intensity within the active separation zone. This effect is particularly significant in systems containing minerals of low electrical conductivity, such as those investigated in this study [40].

Instead of facilitating charge transfer as in conductive media, this condition attenuates the electrostatic coupling between the applied field and free-moving particles. As a result, separator performance progressively deteriorates with increasing particle accumulation. Excessive buildup, commonly associated with extended particle residence times, can overload electrode surfaces and eventually lead to particle detachment driven by system flow dynamics.

2. Materials and Methods

2.1. Sample Preparation

The mineral sample (500 kg) used in the triboelectrostatic separation tests was taken directly from the beneficiation plant of UMA [9,10]. According to Figure 1b, it represents the non-magnetic product (concentrate) yielded by the low-intensity magnetic separator that operates in the industrial plant. The mineral sample was submitted to dry screening through a sieve of 0.65 mm opening mesh, whose oversize (+0.65 mm) was ground and mixed with the undersize (−0.65 mm), aiming at yielding a final product exhibiting P₉₀ $\approx$ 0.6 mm. The product of the sample preparation (screening plus grinding) was named “Testing Material” (TM), because it fed the pilot tests. This sample preparation process allowed us to assume that the particle shapes are approximately similar, reducing the effect of the particle size shape as a varying factor in the performance among the tests conducted. After preparation, the TM was homogenized in a Chevron pile, from which aliquots of 10 kg were taken for drying, prior to triboelectrostatic separation tests. Drying was carried out in a laboratory oven (at 100 °C) endowed with an air exhaustion system. The oven was loaded with 120 kg of material per batch, evenly distributed across 12 trays. After 8 h of drying, the trays were removed and the contents sealed in plastic bags to prevent moisture absorption. The samples were then allowed to cool to ambient temperature prior to testing, in order to eliminate any potential influence of elevated temperatures (>40 °C) on separation performance. For each experiment, the bags were opened immediately before feeding the TBS. The surface humidity of the TM was controlled by an empirical procedure developed by the company STET, the TBS supplier [41]. It consists of inserting the probe HM46, manufactured by Vaisala, into the plastic bag containing dried TM to measure the moisture in the air which fills the existing voids among the particles. Accordingly, if the air moisture between the particles was higher than 1%, the sample of TM was dried again for a further eight hours.

2.2. Sample and Test Products Characterization

The particle size distribution of the TM was determined by wet screening. The chemical composition of the TM and the products generated during the separation tests (concentrate and tailings), including P₂O₅, CaO, SiO₂, Al₂O₃, Fe₂O₃, MgO, TiO₂, and K₂O, were analyzed by X-ray fluorescence (XRF) using the Zetium XRF Spectrometer (Malvern Panalytical, Malvern, UK) with fused bead preparation. Due to the presence of OH-bearing phyllosilicates, the volatile matter content, expressed as loss on ignition (LOI), was determined gravimetrically after roasting the samples at 1020 °C for 2 h.

The mineralogical composition of the TM was assessed by X-ray diffraction (XRD) using the powder method with an Empyrean diffractometer (Malvern Panalytical, Malvern, UK). Crystalline phases were identified by comparing the acquired diffractograms with the PDF-2 database from the International Centre for Diffraction Data (ICDD) and the Inorganic Crystal Structure Database (ICSD). Quantitative mineralogical analysis was conducted via the Rietveld refinement method.

The specific gravity of the TM (2.988 kg/m³) was measured using a pycnometer. The degree of apatite liberation in the TM was evaluated using a mineral liberation analyzer (MLA, FEI), supported by a Quanta 650 FEG scanning electron microscope (SEM) equipped with an energy-dispersive spectroscopy (EDS) system (Bruker).

2.3. Triboelectrostatic Belt Separator (TBS): Description and Mode of Operation

The TBS used in this study was the M6c model (ST Equipment & Technology, Needham, MA, USA), based on the technology described in the patent by David R. Whitlock [42,43], which proposes a continuous triboelectric separation system. In this system, particles of different mineralogical nature acquire charges through contact (triboelectrification) and are subsequently separated under the influence of an electric field established between parallel electrodes. The separator employs a conveyor belt made of dielectric material, providing high mechanical strength, low abrasiveness, and good thermal stability. The belt used in this study was thin, mesh-like, and porous, enabling the simultaneous transport of particles with opposite charges in opposite directions within the inter-electrode space.

Pilot-scale experiments were conducted using the M6c TBS, as illustrated in Figure 4a. Triboelectric charging occurs immediately after the TM is introduced into the feeder, where vigorous agitation ensures efficient contact between particles (Figure 2). The equipment features three feed ports (Figure 4b), although only one is active during operation.

Separation within the M6c TBS is accomplished by 6.1 m long electrode panels positioned at the top and bottom of the machine. Positively charged apatite particles are attracted to the negatively charged electrode, whereas negatively charged phyllosilicate particles are drawn to the positively charged electrode. The equipment allows electrode polarity to be reversed; both configurations were tested in this study. A conveyor belt running between the electrode panels sweeps the separated particles into two distinct hoppers (E1 and E2, Figure 4b) located at either end of the separator. During operation, positively charged apatite particles were consistently directed toward the negatively charged electrode panel and transported by the belt to End 1 (E1), with belt direction adjusted as required. Conversely, negatively charged gangue particles were attracted to the positively charged electrode panel and transported to End 2 (E2).

The operation of the M6c TBS is governed by four primary independent variables, listed in

Table 4. Independent variables of TBS.

Variables	Units	Typical Range
Top electrode polarity	-	Positive/negative
Electrode voltage	$\mathrm{k}\mathrm{V}$	$\pm$6 (*)
Belt speed (S)	$\mathrm{m}/\mathrm{s}$	4.6–19.8
Feed port	-	1, 2, and 3
Electrode gap (G)	$\mathrm{m}$	1.0–1.5 (×10−2)
Mass feed rate of solids (F)	$\mathrm{k}\mathrm{g}/\mathrm{s}$	0.13–1.25

(*) Constant due to results from exploratory studies.

: electrode voltage, electrode gap (G), belt speed (S), and mass feed rate of solids (F). These variables are considered “independent” because they are directly adjusted via the equipment control panel. In the early beginning of the pilot tests, the operability of the equipment with the phosphate ore was used to determine narrower working ranges for the magnitude of the equipment’s operational variables. Further optimization experiments were carried out exploring the magnitude of the operational variables within the boundaries previously and empirically determined. Once the operational settings were established, seven response parameters, called secondary variables in

Table 5. Secondary variables of TBS.

Secondary Variables	Units	Assessment
Cross-sectional area (A)	${\mathrm{m}}^2$	Based on electrode gap (G)
Solids volumetric flowrate (Q)	${\mathrm{m}}^3/\mathrm{s}$	Equation (2)
Solids flux velocity (V)	$\mathrm{m}/\mathrm{s}$	Equation (3)
Total flux velocity ($V_T$)	$\mathrm{m}/\mathrm{s}$	Equation (4)
Residence time (T)	s	Equation (5)
Electric current on the electrodes (I)	mA	Current meter (*)

(*) Informed by the control panel.

, were monitored to interpret the experimental results: cross-sectional area (A) of the separation chamber, solids volumetric flow rate (Q), solids flux velocity (V), which is the velocity of the particle stream derived from the solids volumetric flow rate (Q) and cross-sectional area (A) of the separator chamber, total solids flux velocity ($V_T$), defined as the sum of the solids flux velocity (V) and the belt speed (S), particle residence time (T) within the separator, and electrical current measured on the top and bottom electrodes (I). The magnitudes of Q, V, and $V_T$, were calculated based on the independent variables (Table 4) using Equations (2)–(4).

The distance (L) traveled by apatite or gangue particles from the feed ports (1, 2, and 3) to the collection points (End 1 and End 2) was determined from the physical dimensions of the M6c TBS and is reported in Table 6. During all tests, apatite was consistently collected at End 2 (E2), while gangue was collected at End 1 (E1). This information enabled calculation of particle residence time (T) using Equation (5). Electrical current through the electrodes was measured with the built-in sensors, with readings displayed on the control panel. Prior to sample introduction, both electrodes displayed a baseline current of 2.0 mA. Upon feeding the sample, current values initially fluctuated before stabilizing, at which point the final readings were recorded.

$$Q={\displaystyle \frac{F}{\delta }}$$

(2)

$$V={\displaystyle \frac{Q}{A}}$$

(3)

$$V_T=V+S$$

(4)

$$T_{E1}={\displaystyle \frac{L_{E1}}{V_T}} \mathrm{and} T_{E2}={\displaystyle \frac{L_{E2}}{V_T}}$$

(5)

where

F = Mass feed rate of solids;

Q = Solids volumetric flowrate;

$\mathsf{\delta }$ = Solids specific gravity (2988 kg/m³);

$\mathrm{V}$ = Solids flux velocity (m/s);

${\mathrm{V}}_{\mathrm{T}}$ = Overall solids flux velocity (m/s);

${\mathrm{T}}_{\mathrm{E}1}$ = Residence time (s) to reach End 1;

${\mathrm{T}}_{\mathrm{E}2}$ = Residence time (s) to reach End 2

${\mathrm{L}}_{\mathrm{E}1}$ = Distance traveled by gangue particles from feed port to reach End 1;

${\mathrm{L}}_{\mathrm{E}2}$ = Distance traveled by apatite particles from feed port to reach End 2.

Table 6. Distance traveled by particles of gangue minerals (L_E1) and apatite (L_E2) from each feed port.

Feed Port	LE1 (m)	LE2 (m)
1	4.58	1.53
2	3.05	3.05
3	1.53	4.58

Table 6. Distance traveled by particles of gangue minerals (L_E1) and apatite (L_E2) from each feed port.

Feed Port	LE1 (m)	LE2 (m)
1	4.58	1.53
2	3.05	3.05
3	1.53	4.58

2.4. Pilot Tests

Pilot-scale tests were performed through 20 experimental runs, each conducted in 10 kg batches of TM. All runs were duplicated under the experimental conditions listed in

Table 7. Operational conditions used to perform pilot tests.

Run	Duty	Independent Variables					Secondary Variables
		Feed Port	Top Electrode Polarity	Feed Rate (kg/s)	Gap (×10−2 m)	Belt Speed (m/s)	Solids Flowrate (×10−4 m3/s)	Total Flux Velocity (m/s)	Time to Reach E2 (s) (*)	Time to Reach E1 (s) (**)	Top Electrode Current (mA)	Bottom Electrode Current (mA)
#1	Rougher	1	Positive	0.56	1.52	18.3	1.9	18.4	0.08	0.25	2.3	2.3
2	Rougher	1	Positive	0.56	1.32	12.2	1.9	12.4	0.12	0.37	1.3	1.3
3	Rougher	1	Positive	0.47	1.21	6.1	1.6	6.3	0.24	0.73	0.6	1.5
4	Rougher	2	Positive	0.44	1.25	8.4	1.5	8.6	0.36	0.36	0.8	0.8
5	Rougher	2	Positive	0.44	1.27	6.1	1.5	6.3	0.49	0.49	0.4	0.4
6	Rougher	2	Positive	0.39	1.14	4.6	1.3	4.8	0.64	0.64	0.5	0.5
7	Rougher	3	Positive	0.39	1.14	6.1	1.3	6.3	0.73	0.24	0.7	0.7
8	Rougher	3	Positive	0.44	1.14	5.2	1.5	5.4	0.85	0.28	0.5	0.6
9	Rougher	1	Negative	0.56	1.08	12.2	1.9	12.5	0.12	0.37	2.4	2.3
10	Rougher	1	Negative	0.56	1.13	9.1	1.9	9.4	0.16	0.49	2.8	1.9
11	Rougher	1	Negative	0.56	1.20	6.1	1.9	6.3	0.24	0.72	2.2	1.4
12	Rougher	3	Negative	0.56	1.14	12.2	1.9	12.5	0.37	0.12	2.7	1.8
13	Rougher	3	Negative	0.56	1.08	6.1	1.9	6.4	0.71	0.24	2.2	1.1
14	Rougher	3	Negative	0.56	1.14	6.1	1.9	6.4	0.72	0.24	2.8	2.1
15	Cleaner	1	Positive	0.22	1.33	15.2	0.7	15.3	0.10	0.30	0.6	0.6
16	Cleaner	1	Positive	0.56	1.27	12.2	1.9	12.4	0.12	0.37	0.5	0.5
17	Cleaner	1	Positive	0.56	1.27	6.1	1.9	6.3	0.24	0.73	1.7	1.8
18	Cleaner	2	Positive	0.56	1.27	12.2	1.9	12.4	0.25	0.25	0.6	0.6
19	Cleaner	2	Positive	0.31	1.27	9.1	1.0	9.3	0.33	0.33	0.8	0.7
20	Cleaner	2	Positive	0.48	1.23	6.1	1.6	8.3	0.48	0.48	1.5	1.5

(*) Concentrate; (**) Tailings.

, and the reported electrical currents for the top and bottom electrodes correspond to the averages of the duplicate measurements. The first 14 runs (run 1 to run 14) explored the independent variables from Table 4 in a rougher stage. Prior to each new experimental run, the M6c TBS was cleaned using 5 kg of fresh TM (“flush”) to remove residues from the previous test. Since none of the rougher concentrates achieved a P₂O₅ grade above 28.5%, six additional runs (run 15 to run 20) were conducted to clean the rougher concentrate. To generate sufficient material for these cleaning runs, the operational conditions of the best-performing rougher run (run 4) were repeated 12 times on 20 kg batches. The resulting rougher concentrates were homogenized in a Chevron pile, from which 10 kg samples were collected for subsequent cleaner-stage experiments.

Each experimental run produced two fractions: concentrate, collected at End 2 (E2), and tailings, collected at End 1 (E1) (Figure 4b). These fractions were weighed and homogenized in Chevron piles, and aliquots were analyzed chemically by X-ray fluorescence. The mass yield (Y) of each test, as well as the metallurgical recovery (R) of target analytes (e.g., P₂O₅, MgO), was calculated using Equations (6) and (7), respectively. The efficiency of apatite/phyllosilicate separation was evaluated using the separation efficiency (E) metric proposed by Shultz [44], as defined in Equation (8).

$$Y=100 {\displaystyle \frac{C}{C+T}}$$

(6)

$$R=100 {\displaystyle \frac{C·c}{(C·c)+(T·t)}}$$

(7)

$$E=R_{apatite}-R_{phyllosilicates}$$

(8)

where

Y = Mass recovery (%)

C = Mass of concentrate (recovered in the electrode E2);

T = Mass of tailings (recovered in the electrode E1);

R = Recovery of P₂O₅ or other key analytes, as MgO;

c = Content of a target analyte (P₂O₅, MgO, etc.) in the concentrate;

t = Content of a target analyte (P₂O₅, MgO, etc.) in the tailings;

E = Shultz efficiency (%).

${\mathrm{R}}_{\mathrm{a}\mathrm{p}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{t}\mathrm{e}}$ = Recovery of apatite, assessed by Equation (8) using the content of P₂O₅;

${\mathrm{R}}_{\mathrm{p}\mathrm{h}\mathrm{y}\mathrm{l}\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{i}\mathrm{l}\mathrm{i}\mathrm{c}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{s}}$ = Recovery of phyllosilicates by Equation (8), using the content of MgO.

3. Results and Discussion

3.1. Sample Characterization

Table 8. Particle size distribution of the testing material (TM).

Size Fractions (mm)	Mass (%)
	Retained	Accumulated
+0.600	9.8	9.8
−0.600 + 0.500	10.3	20.1
−0.500 + 0.300	14.6	34.7
−0.300 + 0.210	17.3	52.0
−0.210 + 0.150	14.5	66.5
−0.150 + 0.074	16.7	83.2
−0.074	16.8	100.0
Total	100.0	-

shows the particle size distribution of the sample tested, confirming that the sample was prepared to have a D₉₀ ≈ 0.6 mm. As shown in

Table 9. Chemical composition of the testing material (TM) by size.

Size Fractions (mm)	Mass (%)		Content of Analytes (%)
	Retained	Accum. (*)	P2O5	CaO	SiO2	Al2O3	Fe2O3	MgO	TiO2	K2O	LOI	CaO:P2O5
+0.30	34.7	34.7	25.1	33.8	15.4	4.35	4.79	6.76	0.45	1.18	4.08	1.35
−0.30 + 0.21	17.3	52.0	26.5	35.5	14.1	3.81	5.01	5.56	0.49	1.03	3.73	1.34
−0.21 + 0.15	14.5	66.5	27.3	36.6	12.9	3.40	5.29	4.95	0.65	0.93	3.51	1.34
−0.15 + 0.074	16.7	83.2	25.0	33.9	15.1	4.01	7.09	5.26	0.51	0.99	4.12	1.36
−0.74	16.8	100.0	11.4	16.3	27.4	8.51	14.6	7.39	0.71	1.34	7.67	1.43
Head (TM)	100.0	-	22.9	31.0	17.4	4.99	6.81	6.50	0.48	1.16	4.62	1.35
Total Calc.	100.0	-	23.3	31.6	16.8	4.76	6.94	6.15	0.54	1.11	4.55	1.35

(*) Mass accumulated above.

, the material used in the pilot tests (TM) is primarily composed of CaO (31.0%) and P₂O₅ (22.9%), resulting in a CaO/P₂O₅ ratio of 1.35, slightly higher than the stoichiometric ratio of apatite (~1.32). Ratios above 1.32 indicate the presence of carbonates (calcite and dolomite), which is particularly evident in the finest size fraction (<0.74 mm). Major chemical contaminants, including SiO₂ (17.4%), Fe₂O₃ (6.81%), MgO (6.50%), and Al₂O₃ (4.99%), together with K₂O (1.16%) and LOI (4.62%), suggest the abundance of phyllosilicates. Qualitative mineralogical analysis (Figure 5) confirms that the TM predominantly consists of apatite and phyllosilicates (vermiculite, vermiculite interstratified with hydromica, and smectite), with minor components including amphibole, carbonates (calcite, dolomite), and plagioclase feldspar.

The mineralogical composition of the sample is illustrated in Figure 5 by the X-ray diffraction (XRD) pattern, which shows distinct peaks corresponding to vermiculite, biotite, amphibole, hydroxylapatite, and others, thereby confirming the composition reported in

Table 10. Mineralogical composition of the testing material (TM) by size.

Mineral Species	Content (%)
	Fraction +0.600 mm	Fraction −0.600 + 0.074 mm
Apatite	61	60
Phyllosilicates	32	30
Feldspar	2.5	2.5
Quartz	0.8	1.2
Pyroxene + amphibole	1.2	2.0
Iron oxides (*)	1.1	1.7
Carbonates (**)	1.1	1.1
Titanium oxides (***)	0.1	0.5
Psilomelane	0.1	0.2
Others	0.1	0.8

(*) Magnetite, hematite, goethite; (**) Calcite, dolomite; (***) Ilmenite, anatase.

. The mineralogical composition of the size fraction between –0.60 mm and +0.078 mm of the TM shown by Table 10 closely reflects its overall chemical composition (Table 9). This fraction is predominantly composed of apatite (60%–61%), followed by phyllosilicates (30%–32%), with minor constituents, including feldspar (≈2.5%), pyroxene/amphibole (1.2%–2.0%), Fe-oxides (1.1%–1.7%), calcite/dolomite (≈1.1%), quartz (0.8%–1.2%), Ti-oxides (0.1%–0.5%), traces of psilomelane (0.1%–0.2%), and other minor phases (0.1%–0.8%).

According to Ferguson’s triboelectric series [24] (apatite >>> tremolite > hydrous silicates > actinolite > pyroxene > feldspar > quartz; Table 2), apatite particles exhibit a much higher tendency to acquire positive charge than the Si-bearing minerals present in the TM (Table 10), indicating a potentially selective separation of apatite from silicates via TES. Preliminary laboratory-scale studies confirmed this trend, motivating the pilot-scale experiments. In contrast, Fe-bearing oxides occupy positions in Ferguson’s [24] and Fraas’ [25] triboelectric series close to that of apatite, suggesting that separation between apatite and oxides may be less selective.

Furthermore, the distribution of key analytes (P₂O₅, CaO, SiO₂, Al₂O₃, Fe₂O₃, MgO, TiO₂, K₂O) among the mineral species in the TM (

Table 11. Partition of the main analytes that make up the testing material (TM).

Minerals	P2O5	CaO	SiO2	Al2O3	Fe2O3	MgO	TiO2	K2O
Apatite	100	97
Phyllosilicates	<1	1	78	85	68	95	56	86
Feldspar			9	13				13
Quartz			7
Other Silicates		1	5	2	4	4		<1
Hematite/magnetite				<1	23	<1	8
Goethite					3
Ilmenite					3		31
Psilomelane
Others	<1	<1	1	1	<1	<1	5

) provides a basis for evaluating separation selectivity. Nearly all P₂O₅ (~100%) and 97% of CaO are contained in apatite, whereas phyllosilicates host 95% of total MgO and 85% of total Al₂O₃. Accordingly, the concentrations of P₂O₅ (representing apatite) and MgO (representing phyllosilicates) in the products of all triboelectrostatic separation tests were used to calculate separation efficiency [39] via Equation (8). Other analytes (SiO₂, TiO₂, Al₂O₃, and Fe₂O₃) were considered only to assess the quality of the resulting concentrates.

The particle size distribution of the TM (Table 8) indicates that 65.3% of the total mass consists of particles smaller than 0.300 mm, which partially falls within the suitable range for processing with the M6c TBS. Scanning electron microscopy (SEM) equipped with WDS, combined with image analysis using the mineral liberation analyzer (MLA), was employed to assess the degree of liberation of apatite and phyllosilicates. Apatite particles exhibit a high degree of liberation (>90%) even in the coarsest size fraction (+0.6 mm) of TM, as illustrated in Figure 6. Although this particle size distribution is not fully optimal for TBS operation, according to the manufacturer’s recommendation (−0.5 mm) [31], it was selected for testing due to the high liberation degree (>90%).

3.2. Triboelectrostatic Separation Test Results

The results from the first 14 experiments (run 1 to run 14) of the pilot campaign are presented in

Table 12. Performance of triboelectrostatic separation experiments carried out in the rougher stage.

Run	Concentrate Composition (%)					Recovery (%)
	P2O5	SiO2	Fe2O3	Al2O3	MgO	Mass	P2O5	MgO
1	24.5	11.0	6.18	3.78	3.75	81.7	88.1	45.6
2	24.7	11.4	6.20	3.86	4.16	80.2	86.9	50.8
3	24.9	10.9	7.33	3.84	3.51	76.9	83.2	51.6
4	25.3	10.9	5.85	3.52	3.37	75.3	85.1	35.8
5	25.8	11.2	5.82	3.60	3.33	68.7	77.1	45.8
6	26.5	10.6	5.61	3.41	4.22	59.3	68.9	39.5
7	27.0	11.0	5.58	3.70	3.83	46.9	55.3	27.9
8	25.6	11.0	5.92	3.54	3.56	65.1	75.4	46.3
9	24.1	9.8	8.08	3.89	3.14	48.0	53.8	38.7
10	22.7	10.5	9.41	4.13	4.04	41.3	44.3	36.8
11	24.0	10.1	8.85	3.76	3.36	12.9	14.1	10.6
12	25.3	9.0	7.62	3.08	2.42	4.5	5.2	2.6
13	26.1	10.4	6.70	3.76	3.67	3.7	4.4	3.1
14	23.7	10.4	8.61	4.11	3.66	2.6	3.0	1.9

. The products collected at End 2 (concentrate) did not achieve P₂O₅ grades above 28.5%, indicating the necessity of a cleaner stage to further enrich the rougher concentrate and meet the specifications established by the phosphate mining company (Fosnor).

3.2.1. Rougher Stage

Results from the rougher experiments were used to investigate the influence of operational variables on the separation process. According to the triboelectric series reported by Ferguson [24] and Fraas [25], apatite particles tend to acquire a positive charge after contact or rubbing with gangue minerals such as silicates and oxides. Consequently, during triboelectric separation, apatite particles are likely attracted to the negatively charged electrode, whereas gangue particles, predominantly Mg-bearing phyllosilicates, are attracted to the positively charged electrode. Figure 7 shows that the highest P₂O₅ recoveries (run 1 to run 8) were achieved when negatively charged particles (gangue) were attracted to the positively charged electrode positioned at the top. The additional lifting effect showed by phyllosilicate particles, compared to apatite, may result from its lower specific gravity (2.727 kg/m³ measured by pycnometer) and platy shape (Table 3). Accordingly, the positive electrode configuration on top was identified as the best condition to operate the M6c TBS for this specific application. The results from run 1 to run 8 follow a second-degree polynomial (r² = 0.95), as expressed by Equation (9). In contrast, runs 9 to 14 exhibited very low metallurgical recoveries and scattered grade-recovery data (Figure 7), which could not be fitted to any curve. These findings discourage the use of the M6c TBS with the top electrode set to negative polarity for this specific application.

$$R=-3.54c^2+169.98 c-1949.68$$

(9)

where

R = Recovery of P₂O₅;

c = P₂O₅ content in the concentrate.

Figure 7. Performance of triboelectrostatic separation tests varying the polarity of the top electrode (positive versus negative) of rougher stage.

Figure 7. Performance of triboelectrostatic separation tests varying the polarity of the top electrode (positive versus negative) of rougher stage.

Accounting for the polynomial expressed by Equation (9), the derivative of P₂O₅ recovery (R) with respect to P₂O₅ grade (c) indicates that a maximum recovery of 90.8% is attained at ${\displaystyle \frac{dR}{dc}}$ = 0, corresponding to a P₂O₅ grade of 24%. However, all experiments (run 1 to run 8) conducted with the top electrode at positive polarity yielded concentrates with P₂O₅ grades above 24% (Table 11). Therefore, the Schultz separation efficiency (E) between apatite and phyllosilicates was used to select the optimal experimental conditions for the rougher stage (Figure 8). The highest separation efficiency (E = 49.3%) was achieved under the conditions of run 4: mass flow rate = 0.4 kg/s, belt speed = 8.4 m/s, electrode gap = 1.25 × 10⁻² m, and feed port = 2. These conditions were adopted for the rougher stage, producing a concentrate with 25.3% P₂O₅, mass recovery (yield) of 75.3%, and P₂O₅ recovery of 85.1% (Table 12).

Based on Table 7 and the graphs in Figure 9, which show the electric current measured at the top electrode (left) and the bottom electrode (right) as a function of apatite residence time up to the collection point E2, it is possible to interpret the system’s behavior under different electrode polarity configurations and understand its relationship with separation efficiency and, consequently, apatite recovery. Prior to introducing the ore into the separator, both electrodes exhibited a stable current of 2.0 mA. Once the process was initiated, the current fluctuated until reaching a new steady state, reflecting the interaction of charged particles with the electric field and their trajectories within the separator.

When the top electrode is configured with positive polarity (▲), and the bottom electrode with negative polarity (▲), a significant decrease in current is observed at both electrodes as the residence time (T_E2) increases. This behavior is associated with the attraction of negatively charged particles (predominantly gangue) to the top electrode and positively charged particles (apatite) to the bottom electrode. As discussed in the background, apatite exhibits higher density than gangue, favoring separation consistent with the balance between electrostatic and gravitational forces. Consequently, particle trajectories become more predictable, and the induced currents stabilize at lower values, indicating improved selectivity and reduced disturbance of the electric field.

Conversely, when the top electrode is negatively polarized (●) and the bottom electrode is positive (■), the current at the top electrode remains above 2.0 mA, with little or no decreasing trend. The bottom electrode shows more erratic behavior, with only a slight tendency toward reduction. This instability suggests that particles do not follow well-defined trajectories, interacting with the electric field in a disordered manner. This compromises selectivity, increases contamination of the collected fraction, and results in poorer apatite separation performance.

The graphs in Figure 9 show that, for an apatite residence time (T_E2) of 0.1 s, the currents remain above initial values, likely due to strong interactions and particle movement inducing current on the electrodes. For longer residence times, electrode currents decrease, indicating particle adherence and formation of a dielectric layer along the electrodes. Furthermore, Figure 10, which depicts P₂O₅ recovery as a function of apatite residence time (T_E2) with a positive top electrode during the rougher stage, demonstrates high recovery rates, indicating that apatite is separated almost immediately upon exposure to the electric field. The residence time of 0.36 s in run 4 corresponds to high P₂O₅ recovery (Figure 10) and the best Schultz separation efficiency (Figure 8), suggesting an optimal residence time for maximal particle adherence (observed as decreasing electrode currents). When residence time exceeds 0.4 s, P₂O₅ recovery decreases, indicating potential electrode overloading and particle detachment due to flow dynamics.

Since run 4 yielded the highest Schultz separation efficiency between apatite and phyllosilicates, it was selected for replication and performed 12 times on 20 kg batches to generate sufficient material for the cleaner experiments. The resulting rougher concentrate presented the chemical composition summarized in

Table 13. Chemical composition of the rougher concentrate yielded by twelve tests reproducing the experimental conditions that characterize run 4.

Analytes	$\overline{\mathit{c}}$ (%)	s	${\mathit{s}}_{\mathbf{\%}}$ (*)
P2O5	25.8	0.4	1.6
SiO2	11.4	0.5	4.8
Al2O3	4.0	0.8	5.0
MgO	3.8	0.3	8.8

(*) $s_{\mathrm{\%}}=100(s/\overline{c}).$

. These values differ slightly from those in Table 12, as they represent the average chemical composition of the analytes $(\overline{c}$) obtained from the 12 repeated rougher experiments conducted under the run 4 conditions. Nevertheless, all $\overline{c}$ values reported in Table 13 display low standard deviations (s) and relative standard deviations ($s_{\%}$) below 10%, indicating high reproducibility of the experimental procedure.

3.2.2. Cleaner Stage

After generating sufficient mass for the cleaner stage tests, the first campaign aimed to replicate the configuration of rougher run 4 while exploring the effects of varying belt speed and mass flow rate. Three configurations were tested (runs 15, 16, and 17) using a constant electrode gap of 1.25 × 10⁻² m and Feed Port 2, with mass flow rates ranging from 0.3 to 0.6 kg/s and belt speeds from 6.1 to 12.2 m/s. The results presented in Figure 11 and

Table 14. Performance of triboelectrostatic separation experiments carried out in the cleaner stage.

Runs	Concentrate Composition (%)					Recovery (%)
	P2O5	SiO2	Fe2O3	Al2O3	MgO	Mass	P2O5	MgO
15	29.0	8.62	4.62	2.96	2.84	66.5	81.4	55.3
16	27.9	10.16	5.43	3.28	2.42	67.3	76.7	41.8
17	27.7	10.34	5.19	3.14	3.06	62.3	71.0	50.2
18	28.9	9.78	4.81	2.99	2.89	72.8	87.6	59.6
19	27.2	9.51	5.10	3.14	2.99	63.5	71.6	55.1
20	31.9	8.89	3.94	2.29	1.64	20.8	33.2	12.9

show that the P₂O₅ grade of the concentrate decreased as the residence time (T_E2) increased. Based on these observations, a second campaign was designed to investigate shorter residence times by using Feed Port 1, reducing the travel distance of apatite particles (runs 18 and 19).

To further maximize the P₂O₅ grade, an extreme strategy was applied in run 20, combining Feed Port 1 with an increased belt speed and a reduced feed rate of 0.2 kg/s. This configuration produced a concentrate with 31.9% P₂O₅ but a low P₂O₅ recovery of only 33.2%. Notably, this run also yielded the lowest R₂O₃ content (Fe₂O₃ + Al₂O₃) observed during the tests, 6.2%. Considering the target grade of >28.5% P₂O₅, two favorable outcomes were achieved in runs 15 and 18, producing concentrates with 29% and 28.9% P₂O₅, respectively, along with high P₂O₅ recoveries of 81.4% and 87.6%, meeting both quality and recovery criteria.

3.2.3. Full Circuit Configuration Rougher/Cleaner

Feeding the M6c triboelectric belt separator (TBS) with a testing material containing 22.9% P₂O₅ and operating the equipment as a rougher concentrator under the conditions of run 4, the cleaner stage was performed to enrich the P₂O₅ grade of the final concentrate above 28.5%. Two different configuration options were considered based on the cleaner-stage results: (i) option 1 utilized the concentrate from run 15, which produced 29.0% P₂O₅ and reduced all major contaminants (SiO₂, Fe₂O₃, Al₂O₃, and MgO), resulting in an overall mass recovery of 50.0% and a P₂O₅ recovery of 69.3%; (ii) option 2 used the product from run 18, yielding a similar concentrate grade of 28.9% P₂O₅ but achieving higher recoveries, with an overall mass recovery of 54.8% and P₂O₅ recovery of 74.5%. The results are summarized in

Table 15. Summary of the triboelectrostatic separation results compared with Fosnor’s results.

Item	Chemical Composition (%)					Overall Recovery (%)
	P2O5	SiO2	Fe2O3	Al2O3	MgO	Mass	P2O5	MgO
Feed	22.9	17.4	6.81	4.99	6.50	-	-	-
Run 4 (*)	25.3	10.9	5.85	3.52	3.37	75.3	85.1	35.8
Run 15 (*)	29.0	8.62	4.62	2.96	2.84	66.5	81.4	55.3
Run 18 (*)	28.9	9.78	4.81	2.99	2.89	72.8	87.6	59.6
Total Option 1(**)	29.0	8.62	4.62	2.96	2.84	50.0	69.3	19.8
Total Option 2 (***)	28.9	9.78	4.81	2.99	2.89	54.8	74.5	21.3
Fosnor (****)	29.5	9.51	4.94	3.13	1.98	36.3	55.7	27.9

(*) Concentrate product; (**) Considering the performance of run 4 and run 15; (***) Considering the performance of run 4 and run 18; (****) Considering results of high-intensity magnetic separators.

.

When compared to results previously reported by Fosnor using two high-intensity magnetic separators (one at 0.5 T and the other >1.0 T), both TBS configurations exhibited improved performance. Option 1 showed increases of 13.7% in mass recovery and 13.6% in P₂O₅ recovery, while option 2 achieved gains of 18.5% and 18.8%, respectively. Furthermore, MgO removal improved by 8.1% in option 1 and 6.6% in option 2, indicating greater selectivity in phyllosilicate removal.

Although the TBS did not substantially enhance the final P₂O₅ grade compared to magnetic separation, slight reductions in R₂O₃ (Fe₂O₃ + Al₂O₃) were observed: 0.49% for option 1 and 0.2% for option 2. Nevertheless, these values remained above the market specification threshold of <5%.

Despite this limitation, the TBS process clearly outperformed conventional magnetic separation in overall recovery and contaminant removal. These results highlight the potential of triboelectric separation as a viable and efficient alternative for processing this type of phosphate ore.

Although the testing material exhibited a high degree of mineral liberation, the particle size employed in this study (−0.6 mm) may not be optimal for TBS processing, which demonstrates its highest industrial performance with finer fractions, typically below −0.5 mm. Future studies should assess TBS performance at finer particle sizes to determine its true potential under ideal operating conditions. Additionally, the application of the TBS could be explored in other stages of the beneficiation circuit, such as treating the magnetic product from the high-intensity dry magnetic separator, with the aim of recovering apatite losses not captured in the current flowsheet.

Further research should investigate the use of chemical modifiers or surfactants to selectively enhance surface charging, thereby increasing triboelectric contrast and separation efficiency. Complementary development of advanced kinetic models that account for mineral liberation, electrostatic behavior, and particle morphology could substantially improve prediction and optimization of TBS performance. Combined with continued efforts in process modeling and control, these approaches may establish the TBS as a more efficient and versatile technology for phosphate ore beneficiation.

4. Conclusions

This study evaluated the technical feasibility and operational performance of a triboelectrostatic belt separator (TBS) for the dry concentration of apatite in phosphate ores rich in phyllosilicates at the pilot scale. Using the M6c TBS, the operational configuration of run 4 proved most effective, producing a rougher stage concentrate with 25.3% P₂O₅ and a P₂O₅ recovery of 85.1%. In the cleaner stage, run 18 yielded concentrates exceeding 28.5% P₂O₅, with a P₂O₅ recovery of up to 74.5%, surpassing the performance of the high-intensity magnetic separators currently employed at Fosnor’s industrial plant.

Unlike most laboratory-scale studies, this investigation demonstrated the practical applicability of the TBS under real pilot-scale conditions using industrial samples, effectively bridging the gap between fundamental research and industrial implementation. The analysis confirmed that key operational variables—electrode voltage, gap, belt speed, and feed rate—significantly influence separation efficiency.

Moreover, the study proposed an approach based on the Shockley–Ramo theorem and Ohm’s law for resistive media to explain the relationship between electrode current, selectivity, and overall TBS performance. An increase in electrode current was attributed to the erratic motion of charged particles within the electrode gap, inducing current on the electrodes, as predicted by the Shockley–Ramo theorem. Conversely, a decrease in measured current was associated with particle immobilization: when particles adhered to the electrode surface, their contribution to the induced current effectively vanished, and the insulating particle layer acted as a dielectric barrier, further limiting current conduction. Overall, the interaction between electrode polarity and particle properties—such as density, resistivity, and triboelectric behavior—was identified as a critical factor in TBS performance.

The repeatability of twelve repetitions of run 4 highlighted the robustness and operational stability of the TBS on the pilot scale. The data also demonstrated that suboptimal operational configurations reduce recovery and selectivity, emphasizing the importance of precise parameter control. Although R₂O₃ contents in the final concentrates remained slightly above commercial limits, the TBS exhibited superior P₂O₅ recovery and MgO removal. This dry separation approach not only reduces water consumption in phosphate beneficiation but also minimizes wet tailings generation, providing environmental and operational advantages for plants in semi-arid regions. Future work should focus on optimizing the separation of finer particles (−0.3 mm) and exploring alternative electrode configurations and surface charge modifiers to enhance process efficiency and broaden industrial applicability.