The E ﬀ ect of Surface Charge on the Separation of Pyrite from Serpentine by Flotation

: Serpentine, a magnesium silicate mineral with positive surface charge in many sulﬁde ores around the world, usually deteriorates the ﬂotation behavior by covering the target mineral surface. In this paper, the e ﬀ ect of surface potential regulation on serpentine ﬂotation was revealed by ﬂocculation experiments, zeta potential measurements, infrared spectrum analysis, and DLVO theoretical calculations. The experimental results of ﬂocculation and sedimentation show that heterogeneous coagulation easily occurs between serpentine and pyrite particles, which reduces the ﬂoatability of pyrite. Reducing the surface potential of serpentine is an e ﬀ ective way to eliminate heterogeneous coagulation between minerals. The key to regulating the surface potential of serpentine is Mg 2 + ion dissolution from the serpentine surface to the liquid phase. Phosphates, especially sodium hexametaphosphate, can enhance Mg 2 + ion dissolution from the serpentine surface to the liquid phase and react with Mg 2 + ions in the liquid phase to form stable soluble complexes. ﬀ + Figure 3b,d that SHMP can also eliminate heterogeneous ﬂocculation between pyrite and serpentine particles. The results of the background rates in pulp solution (53.47% vs. 53.15%) and the numbers of serpentine particles (1188 vs. 1147) reveal that the dispersion of after the removal of from serpentine and addition of SHMP is similar.


Introduction
Serpentine is one of the most common magnesium-bearing silicate minerals in nickel/copper sulphide, rhodochrosite, and manganese hydrate ores [1][2][3]. When serpentine dissociates, the magnesia-oxygen octahedron layer breaks, and the ionic magnesium-oxygen bond is exposed on the surface [4]. It has strong hydration and good natural hydrophilicity. The surface hydroxyls of fine serpentine dissolve preferentially in an aqueous solution, and Mg 2+ ions remain on the serpentine surface, which makes the surface charge of serpentine stronger and its point of zero charge around a pH of 10.0 [5,6]. In weak alkaline pulp, the surface of sulfide minerals is generally negatively charged, and fine serpentine with a positive charge is adsorbed on the sulfide mineral surface with negative charge under the action of an electrostatic adsorption force, which makes the difference in the hydrophobicity and surface properties of sulfide minerals significantly reduced and deteriorates the flotation behavior of sulfide minerals [6,7]. In view of this property of serpentine, some researchers have used serpentine as an inhibitor for the flotation separation of galena and pyrite [8][9][10].
The cover of fine serpentine particles on the surface of target minerals is the main factor affecting the flotation of sulfide ores by serpentine, which is mainly caused by electrostatic interaction [6,9]. Reducing the surface potential of serpentine and making it better dispersed in solution is the key factor to solving the problem of sulfide mineral flotation containing serpentine [11,12]. In laboratory studies, nickel pyrite was separated from serpentine by magnetic separation after encapsulating the target mineral with fine magnetic particles [13][14][15][16][17]. These studies were only conducted in the laboratory stage, and there are many other factors that need to be considered in magnetic package

Materials and Reagents
The pyrite was obtained from Yunfu, Guangdong Province, China, and the serpentine was obtained from Donghai, Jiangsu Province China. The samples were first removed by hand and then crushed in a ceramic ball 72 mill. The time was generally controlled in 3-5 min. After grinding, 200 mesh and 400 mesh sieves 73 were used for screening, more than 200 mesh sieves were returned to ceramic pots, less than 400 mesh 74 sieves were used as crushed products for zeta potential and FTIR spectroscopy testing, and 200-400 75 mesh sieves were used as flotation raw ore storage. The crystal phase composition was determined by X-ray diffraction (Germany Bruker-axes D8 Advance, Karlsruhe, Germany), the detailed processes and methods for which were previously described by Xiao [22,23]. The chemical compositions of pyrite and serpentine were determined by wavelength-dispersive X-ray fluorescence (XRF) using an S4 Pioneer spectrometer (PANalytical B.V., Almelo, The Netherlands). The results of XRD and XRF are shown in Figure 1 and Table 1, respectively. The results demonstrate that the purities of pyrite and serpentine in the samples were calculated to be 93.9% and 83.6%. The serpentine single mineral was stirred in hydrochloric acid of 2 mol/L for 30 min at room temperature, filtered, washed three times with ultra-pure water, and dried in vacuum at room temperature to obtain the demagnesium serpentine sample. Trisodium phosphonate (TSP), sodium trimetaphosphate (STMP), and sodium hexametaphosphate (SHMP) of analytically pure grade were obtained from Sinopharm Chemical Reagent Co., Ltd (Shanghai, China). Ultra-pure water with a conductivity of 18.2 MΩ·cm was obtained from a USF-ELGA Maxima water purification system [24,25].

Micro-Flotation TESTS
For the flotation test, we used a 40 mL XFG trough-type flotation machine. An ore sample weighing 2 g was placed in the flotation cell for each test. Then, a certain concentration of flotation reagent was added and stirred for 5 min. The pH value of the slurry was measured using a precision pH meter. The flotation process was carried out by manual scraping, and the flotation time was 5 min. The obtained foam product and the product in the tank were dried and weighed, and the yield was calculated. For the single-mineral test, we took the flotation recovery rate to be equal to the yield; in the artificial mixed ore flotation test, the flotation recovery rate was calculated after chemical analysis of each product.

Settlement Experiment
Turbidity was used to characterize the dispersion and aggregation of mineral particles. The pyrite sample with a size range between 38 to 74 μm and the serpentine sample with 104 a size of less than 10 μm were weighed to 1g, respectively, flotation reagents were added and stirred for 5 min, and the mix was then deposited in a 100 mL settling cylinder for 3 min. A volume of 20 mL of the upper suspension was extracted and put into a sample bottle. The turbidity was measured using a WGZ-3(3A) scattering light turbidimeter (Xinrui Instruments, Shanghai, China). When the turbidity is high, there are more mineral particles in the upper suspension of the sedimentation system, and the pulp is in a dispersed state. On the contrary, when the turbidity is low, the mineral particles in the upper suspension are less numerous, and the pulp is in an aggregated state.

Microscopic Observation
A quantity of 2 g of artificial mixed ore sample (serpentine/pyrite = 1:9) was weighed, and flotation reagent of a certain concentration was added and stirred for 5 min. A small amount of slurry droplets was removed by a needle tube under stirring conditions and placed on a slide to observe and photograph the dispersion of minerals under an Olympus CX31 transmission light microscope (Olympus Corporation, Tokyo, Japan). The image head captured the observed electronic image. Image-Pro Plus graphics software was used to analyze and process the obtained microphotographs, and the background rate of the microphotographs and the number of serpentine particles were obtained. The background rate represents the percentage of background area without particles in the image, and the larger the background rate, the smaller the proportion of mineral particles in the area, as the worse the dispersion; conversely, the smaller the background rate, the better the dispersion of mixed minerals [26,27]. The number of particles represents the number of mineral particles with obvious boundaries in the image. The smaller the number of particles, the worse the dispersion of the mixed ore. The more particles there are, the better the dispersion of the mixed ore.

Micro-Flotation TESTS
For the flotation test, we used a 40 mL XFG trough-type flotation machine. An ore sample weighing 2 g was placed in the flotation cell for each test. Then, a certain concentration of flotation reagent was added and stirred for 5 min. The pH value of the slurry was measured using a precision pH meter. The flotation process was carried out by manual scraping, and the flotation time was 5 min. The obtained foam product and the product in the tank were dried and weighed, and the yield was calculated. For the single-mineral test, we took the flotation recovery rate to be equal to the yield; in the artificial mixed ore flotation test, the flotation recovery rate was calculated after chemical analysis of each product.

Settlement Experiment
Turbidity was used to characterize the dispersion and aggregation of mineral particles. The pyrite sample with a size range between 38 to 74 µm and the serpentine sample with 104 a size of less than 10 µm were weighed to 1 g, respectively, flotation reagents were added and stirred for 5 min, and the mix was then deposited in a 100 mL settling cylinder for 3 min. A volume of 20 mL of the upper suspension was extracted and put into a sample bottle. The turbidity was measured using a WGZ-3(3A) scattering light turbidimeter (Xinrui Instruments, Shanghai, China). When the turbidity is high, there are more mineral particles in the upper suspension of the sedimentation system, and the pulp is in a dispersed state. On the contrary, when the turbidity is low, the mineral particles in the upper suspension are less numerous, and the pulp is in an aggregated state.

Microscopic Observation
A quantity of 2 g of artificial mixed ore sample (serpentine/pyrite = 1:9) was weighed, and flotation reagent of a certain concentration was added and stirred for 5 min. A small amount of slurry droplets was removed by a needle tube under stirring conditions and placed on a slide to observe and photograph the dispersion of minerals under an Olympus CX31 transmission light microscope (Olympus Corporation, Tokyo, Japan). The image head captured the observed electronic image. Image-Pro Plus graphics software was used to analyze and process the obtained microphotographs, and the background rate of the microphotographs and the number of serpentine particles were obtained. The background rate represents the percentage of background area without particles in the image, and the larger the background rate, the smaller the proportion of mineral particles in the area, as the worse the dispersion; conversely, the smaller the background rate, the better the dispersion of mixed minerals [26,27]. The number of particles represents the number of mineral particles with obvious boundaries in the image. The smaller the number of particles, the worse the dispersion of the mixed ore. The more particles there are, the better the dispersion of the mixed ore.

Zeta Potential Measurements
The zeta potential of mineral surfaces was measured using a ZetaPALS instrument (Brookhaven Instruments, Brookhaven, MS, USA). The samples were ground to less than 5 µm. The suspension was prepared by adding 50 mg of pure minerals to 5 mL of distilled, deionized water containing 10 −3 mol/L KCl as a supporting electrolyte. The resulting suspension was conditioned for 15 min during which the suspension pH was measured. The pH was adjusted using either NaOH or H 2 SO 4 over the pH range of 2-12.

FT-IR Spectroscopy
We weighed 2 g of the single-mineral sample with a particle size of less than 2 µm, added a certain concentration of flotation agent, stirred well, and let it stand for 30 min. After the mineral was completely settled, the supernatant liquid was sucked out using a pipette. The mineral was thoroughly washed with distilled water and dried naturally after solid-liquid separation. Infrared spectroscopy was carried out using a Nicolet FTIR-740 Fourier-transform infrared spectrometer (Nicolet Instrument, Waltham, MA, USA). Figure 2 shows the effect of Mg 2+ ion dissolution on the aggregation and dispersion behavior of serpentine and pyrite. The results revealed that the dispersion of demagnesium serpentine and pyrite was improved significantly at pH > 7, and the heterogeneous coagulation between particles was weakened. When Mg 2+ ions were added, the dispersion of demagnesium serpentine and pyrite decreased at pH > 8, the mineral particles re-aggregated, and the dispersion decreased. This indicated that the dispersion of serpentine and pyrite requires not only the removal of magnesium from the serpentine surface, but also the dissolution of Mg 2+ ions from the pulp system.

Zeta Potential Measurements
The zeta potential of mineral surfaces was measured using a ZetaPALS instrument (Brookhaven Instruments, Brookhaven, MS, USA). The samples were ground to less than 5 μm. The suspension was prepared by adding 50 mg of pure minerals to 5 mL of distilled, deionized water containing 10 −3 mol/L KCl as a supporting electrolyte. The resulting suspension was conditioned for 15 min during which the suspension pH was measured. The pH was adjusted using either NaOH or H2SO4 over the pH range of 2-12.

FT-IR Spectroscopy
We weighed 2 g of the single-mineral sample with a particle size of less than 2 μm, added a certain concentration of flotation agent, stirred well, and let it stand for 30 min. After the mineral was completely settled, the supernatant liquid was sucked out using a pipette. The mineral was thoroughly washed with distilled water and dried naturally after solid-liquid separation. Infrared spectroscopy was carried out using a Nicolet FTIR-740 Fourier-transform infrared spectrometer (Nicolet Instrument, Waltham, USA). Figure 2 shows the effect of Mg 2+ ion dissolution on the aggregation and dispersion behavior of serpentine and pyrite. The results revealed that the dispersion of demagnesium serpentine and pyrite was improved significantly at pH > 7, and the heterogeneous coagulation between particles was weakened. When Mg 2+ ions were added, the dispersion of demagnesium serpentine and pyrite decreased at pH > 8, the mineral particles re-aggregated, and the dispersion decreased. This indicated that the dispersion of serpentine and pyrite requires not only the removal of magnesium from the serpentine surface, but also the dissolution of Mg 2+ ions from the pulp system. The dispersion of serpentine and pyrite in pulp can be observed directly by a microscope [24,28]. Figure 3 shows microscopic photos of serpentine and pyrite in the pulp. In the microscopic photos, the large black particles are pyrite and the small light grains are serpentine.  The dispersion of serpentine and pyrite in pulp can be observed directly by a microscope [24,28]. Figure 3 shows microscopic photos of serpentine and pyrite in the pulp. In the microscopic photos, the large black particles are pyrite and the small light grains are serpentine.

Effect of Mg 2+ Ions on the Heterogeneous Flocculation Behavior of Pyrite and Serpentine
After removing magnesium from serpentine, the background rate of the pulp solution decreased from 62.32% to 53.47%, and the number of serpentine particles increased from 592 to 1188, which indicates that Mg 2+ ions on the serpentine surface seriously affected the dispersion effect between mineral particles. After adding Mg 2+ ions to the demagnesium serpentine pulp, the background rate obviously increased (from 53.47% to 58.15%), and the number of serpentine particles decreased to 841, which indicates that the dispersion effect between mineral particles could also be affected by Mg 2+ ions in solution. Comparison with Figure 3b,d shows that SHMP can also eliminate heterogeneous flocculation between pyrite and serpentine particles. The results of the background rates in pulp solution (53.47% vs. 53.15%) and the numbers of serpentine particles (1188 vs. 1147) reveal that the dispersion of solutions after the removal of magnesium from serpentine and addition of SHMP is similar.

Sample
Background Rate (%) There was obvious heterogeneous agglomeration between the pyrite and serpentine particles before magnesium removal (in Figure 3a). After magnesium was removed from the serpentine surface (in Figure 3b), the dispersion of pyrite and serpentine in the pulp became better, and the heterogeneous agglomeration between particles was weakened. When Mg 2+ ions were added to the pulp of pyrite and demagnesium serpentine (in Figure 3c), the particles of pyrite and serpentine re-aggregated, and the dispersion of the pulp became poorer. Table 2 shows the background rates of the microphotographs and the numbers of serpentine particles under various experimental conditions. The smaller the number of particles, the worse the dispersion of serpentine and pyrite, and the more numerous the particles, the better the dispersion. After removing magnesium from serpentine, the background rate of the pulp solution decreased from 62.32% to 53.47%, and the number of serpentine particles increased from 592 to 1188, which indicates that Mg 2+ ions on the serpentine surface seriously affected the dispersion effect between mineral particles. After adding Mg 2+ ions to the demagnesium serpentine pulp, the background rate obviously increased (from 53.47% to 58.15%), and the number of serpentine particles decreased to 841, which indicates that the dispersion effect between mineral particles could also be affected by Mg 2+ ions in solution. Comparison with Figure 3b,d shows that SHMP can also eliminate heterogeneous flocculation between pyrite and serpentine particles. The results of the background rates in pulp solution (53.47% vs. 53.15%) and the numbers of serpentine particles (1188 vs. 1147) reveal that the dispersion of solutions after the removal of magnesium from serpentine and addition of SHMP is similar. Reducing the density of localized ions in the double layer is an effective way to control the surface charge properties of particles [26,29]. As Mg 2+ ions are positioning ions in a serpentine diffused double layer, reducing the concentration of Mg 2+ on the serpentine surface can reduce the positive charge density of the positioning ions in the double layer, thus reducing the surface potential of serpentine. Figure 4 shows the effect of Mg 2+ ion dissolution on the zeta potential of the serpentine surface. After removing magnesium from the serpentine, the zeta potential of the serpentine shifted negatively, and the isoelectric point (IEP) decreased from a pH of 10.0 to 5.4. When Mg 2+ ions (1 × 10 −3 mol·L −1 ) were added to the demagnesium serpentine slurry, the zeta potentials of the serpentine surface decreased slightly at a pH value below 8, and were maintained at about +20 mV at a pH value above 8. This suggests that the surface potential could recover a high positive value by the addition of Mg 2+ ions. Therefore, reducing the zeta potential of serpentine requires not only the dissolution of Mg 2+ ions from the mineral surface, but also the removal of dissolved Mg 2+ ions from the slurry system. Reducing the density of localized ions in the double layer is an effective way to control the surface charge properties of particles [26,29]. As Mg 2+ ions are positioning ions in a serpentine diffused double layer, reducing the concentration of Mg 2+ on the serpentine surface can reduce the positive charge density of the positioning ions in the double layer, thus reducing the surface potential of serpentine. Figure 4 shows the effect of Mg 2+ ion dissolution on the zeta potential of the serpentine surface. After removing magnesium from the serpentine, the zeta potential of the serpentine shifted negatively, and the isoelectric point (IEP) decreased from a pH of 10.0 to 5.4. When Mg 2+ ions (1 × 10 −3 mol·L −1 ) were added to the demagnesium serpentine slurry, the zeta potentials of the serpentine surface decreased slightly at a pH value below 8, and were maintained at about +20 mV at a pH value above 8. This suggests that the surface potential could recover a high positive value by the addition of Mg 2+ ions. Therefore, reducing the zeta potential of serpentine requires not only the dissolution of Mg 2+ ions from the mineral surface, but also the removal of dissolved Mg 2+ ions from the slurry system.

Phosphate Complexation on Surface Potentials of Serpentine
The complexation of Mg 2+ ions on the serpentine surface can effectively regulate its surface potential and eliminate the heterogeneous condensation between serpentine and other mineral particles. On this basis, the adsorption of anion modifier on the serpentine surface further reduces its surface potential and strengthens the regulation of the surface charge properties of serpentine [9,30]. Figure 5 shows the effect of three phosphates on the zeta potential of the demagnesium serpentine surface. In Figure 5a, trisodium phosphonate (TSP) showed little effect on the zeta potential of the demagnesium serpentine surface. Sodium hexametaphosphate (STMP) and sodium trimetaphosphate were able to significantly reduce the zeta potential of the demagnesium serpentine surface, especially SHMP. Figure 5b shows that the zeta potential of demagnesium serpentine

Phosphate Complexation on Surface Potentials of Serpentine
The complexation of Mg 2+ ions on the serpentine surface can effectively regulate its surface potential and eliminate the heterogeneous condensation between serpentine and other mineral particles. On this basis, the adsorption of anion modifier on the serpentine surface further reduces its surface potential and strengthens the regulation of the surface charge properties of serpentine [9,30]. Figure 5 shows the effect of three phosphates on the zeta potential of the demagnesium serpentine surface. In Figure 5a, trisodium phosphonate (TSP) showed little effect on the zeta potential of the demagnesium serpentine surface. Sodium hexametaphosphate (STMP) and sodium trimetaphosphate were able to significantly reduce the zeta potential of the demagnesium serpentine surface, especially SHMP. Figure 5b shows that the zeta potential of demagnesium serpentine gradually decreased with increasing phosphate dosage, and finally tended to be stable. The ability of the three phosphates to reduce the zeta potential of demagnesium serpentine was in the order of SHMP > STMP > TSP. Minerals 7 of 14 gradually decreased with increasing phosphate dosage, and finally tended to be stable. The ability of the three phosphates to reduce the zeta potential of demagnesium serpentine was in the order of SHMP > STMP > TSP.

FT-IR Spectra of Phosphate Adsorption on the Serpentine Surface
The presence of phosphate adsorption on the serpentine surface is shown in Figure 6, which presents the infrared spectra of the phosphate and serpentine before and after phosphate adsorption. In the infrared spectrum of serpentine, the adsorption peak at 3686.0 cm −1 corresponds to the stretching vibration peak of hydroxyl groups in the serpentine magnesia-oxygen octahedron [5]. The broad and strong adsorption peak at 975. 8  In the infrared spectrum of TSP (black line in Figure 6a), the asymmetric stretching vibration and the bending vibration peaks of P-O in PO3 correspond to 1009.7 and 551.6 cm −1 , respectively. After TSP was adsorbed on the serpentine surface (red line in Figure 6a), there was no obvious TSP characteristic absorption peak, but the position of the Mg-O characteristic absorption peak in serpentine shifted considerably (offset by more than 10 cm −1 ). In the infrared spectrum of SHMP (black line in Figure 6b), the peaks at 1279.9, 1096.8, and 875 cm −1 correspond to the P-O stretching vibration, and the P-O bending vibration peak is located at 515.4 cm −1 [31]. In the infrared spectrum of STMP (black line in Figure 6c), the peaks at 1297.0 and 997.9cm −1 correspond to the P-O stretching vibration, and the P-O bending vibration peak is located at 519.2 cm −1 . The infrared spectra of the serpentine surface after SHMP and STMP adsorption are similar to that of TSP adsorption. Similarly, the P-O characteristic peak was not found, and the Mg-O characteristic peak in serpentine also showed a large displacement. This indicated that all three phosphates were chemically adsorbed on the serpentine surface, but their adsorption capacity was very little.

FT-IR Spectra of Phosphate Adsorption on the Serpentine Surface
The presence of phosphate adsorption on the serpentine surface is shown in Figure 6, which presents the infrared spectra of the phosphate and serpentine before and after phosphate adsorption. In the infrared spectrum of serpentine, the adsorption peak at 3686.0 cm −1 corresponds to the stretching vibration peak of hydroxyl groups in the serpentine magnesia-oxygen octahedron [5]. The broad and strong adsorption peak at 975. 8  In the infrared spectrum of TSP (black line in Figure 6a), the asymmetric stretching vibration and the bending vibration peaks of P-O in PO 3 correspond to 1009.7 and 551.6 cm −1 , respectively. After TSP was adsorbed on the serpentine surface (red line in Figure 6a), there was no obvious TSP characteristic absorption peak, but the position of the Mg-O characteristic absorption peak in serpentine shifted considerably (offset by more than 10 cm −1 ). In the infrared spectrum of SHMP (black line in Figure 6b), the peaks at 1279.9, 1096.8, and 875 cm −1 correspond to the P-O stretching vibration, and the P-O bending vibration peak is located at 515.4 cm −1 [31]. In the infrared spectrum of STMP (black line in Figure 6c), the peaks at 1297.0 and 997.9cm −1 correspond to the P-O stretching vibration, and the P-O bending vibration peak is located at 519.2 cm −1 . The infrared spectra of the serpentine surface after SHMP and STMP adsorption are similar to that of TSP adsorption. Similarly, the P-O characteristic peak was not found, and the Mg-O characteristic peak in serpentine also showed a large displacement. This indicated that all three phosphates were chemically adsorbed on the serpentine surface, but their adsorption capacity was very little.

Effect of Surface Charge on the Interaction Energy between Mineral Particles
According to the classical DLVO theory, the total potential energy VT of particle interaction in heterogeneous mineral water-based suspensions is derived by Equation (1) [32]: where A is the effective Hamaker constant of Minerals 1 and 2 in a water medium, R1 is the radius of a Mineral 1 spherical particle, R2 is the radius of a Mineral 2 spherical particle, and H is the distance between Mineral 1 and Mineral 2 particles. The expression of A is [33] = + where VW is the Van der Waals interaction energy and VE is the electrostatic interaction energy. The Van der Waals interaction between spherical particles can be governed by Equation (2): where A is the effective Hamaker constant of Minerals 1 and 2 in a water medium, R1 is the radius of a Mineral 1 spherical particle, R2 is the radius of a Mineral 2 spherical particle, and H is the distance between Mineral 1 and Mineral 2 particles. The expression of A is [33] where A11 and A22 are the Hammaker constants of Mineral 1 and 2 in vacuum, respectively, and A33 is the Hammaker constant of water in vacuum (A33 = 3.7 × 10 −20 J) [34]. The expression of the electrostatic interaction energy VE between the particles of Mineral 1 and Mineral 2 is as follows:

Effect of Surface Charge on the Interaction Energy between Mineral Particles
According to the classical DLVO theory, the total potential energy V T of particle interaction in heterogeneous mineral water-based suspensions is derived by Equation (1) [32]: where V W is the Van der Waals interaction energy and V E is the electrostatic interaction energy. The Van der Waals interaction between spherical particles can be governed by Equation (2): where A is the effective Hamaker constant of Minerals 1 and 2 in a water medium, R 1 is the radius of a Mineral 1 spherical particle, R 2 is the radius of a Mineral 2 spherical particle, and H is the distance between Mineral 1 and Mineral 2 particles. The expression of A is [33] where A 11 and A 22 are the Hammaker constants of Mineral 1 and 2 in vacuum, respectively, and A 33 is the Hammaker constant of water in vacuum (A 33 = 3.7 × 10 −20 J) [34]. The expression of the electrostatic interaction energy V E between the particles of Mineral 1 and Mineral 2 is as follows: where ε 0 is the absolute dielectric constant in vacuum (ε 0 = 8.854 × 10 −12 ); ε r is the dielectric constant of the dispersing medium (water, ε r = 78.5 C 2 ·J −1 ·m −1 ); ψ 1 and ψ 2 are the surface potentials of Mineral 1 and Mineral 2 particles, respectively; ζ is the zeta potential of the mineral solid-liquid interface; x is the distance from the surface of charged mineral particles to the slip surface (x = 5 × 10 −10 m); and κ −1 is the Debye length, representing the thickness of the double layer. ψ and κ are obtained by Equations (5) and (6), respectively: where e is the electron charge (e = 1.602 × 10 −19 C); N A is the Avogadro constant (N A = 6.02 × 10 23 ); I is the ionic strength of the solution (I = 1 × 10 −3 mol/L); K B is the Boltzmann constant (K B = 1.38 × 10 −23 J/K); and T is the thermodynamic temperature (T = 298 K).
In the aqueous suspension system of serpentine and pyrite, A 11 = 9.7 × 10 −20 J, A 22 = 1.2 × 10 −19 J [35], R 1 = 3.09 × 10 −6 m, and R 2 = 2.67 × 10 −5 m. By synthesizing Equations (1) -(6), the following can be concluded: The flotation of copper-nickel sulphide ores is generally carried out under conditions of a weak alkaline pH. For a pulp pH of 9.0, the results of the zeta potentials of the serpentine and pyrite surfaces are shown in Table 3. According to the results in Table 3, the relationship between the total potential energy of interaction between serpentine and pyrite particles in the water medium and the distance between particles is shown in Figure 7. When the pH is 9, the interaction energy between serpentine and pyrite particles is negative. The interaction force between serpentine and pyrite particles shows mutual attraction (in Curve 1). With decreasing particle spacing, the attraction gradually increases and heterogeneous agglomeration easily occurs. The interaction energy between serpentine and pyrite becomes positive, the interaction force between serpentine and pyrite particles is mutually exclusive, and the dispersion between serpentine and pyrite particles is improved (in Curve 2). When sodium hexametaphosphate is added, the interaction energy between serpentine and pyrite is positive, and the interaction force between particles is mutually exclusive. At this time, the exclusion energy between particles is greater than that between serpentine and pyrite after magnesium removal (in Curve 3). Serpentine and pyrite can maintain a better dispersion state in the system. Minerals 10 of 14 Figure 7. Relation between total potential energy of interaction between serpentine and pyrite particles in aqueous media and particle spacing.

Effect of Surface Potential Regulation on the Flotation Separation of Pyrite and Serpentine
The effect of surface potential regulation on the flotation separation of pyrite and serpentine was investigated by a flotation test of artificial mixed ore, as shown in Figure 8. In Figure 8a, we can see that the single mineral of pyrite had good floatability at pH < 10, and the flotation recovery was able to reach more than 90%. The addition of serpentine significantly reduced the flotation recovery of pyrite. After removing magnesium from the serpentine, the flotation recovery of pyrite in the artificial mixed ore increased, and the floatability of pyrite was restored. After removing magnesium from the serpentine and adding Mg 2+ ions, the flotation recovery of pyrite in the artificial mixed ore decreased rapidly at pH > 8, and the floatability of pyrite was affected by serpentine again. As the surface charge properties of serpentine can be significantly reduced by chain polyphosphate, the effects of sodium pyrophosphate, sodium tripolyphosphate, and sodium hexametaphosphate on the flotation separation of serpentine and pyrite were investigated, as shown in Figure 8b. The effect of chain polyphosphates on the flotation of artificial mixed ores was assessed under different pH values. It can be seen from the graph that all three kinds of polyphosphates were able to improve the flotation recovery of pyrite in the mixed ore, restore the floatability of pyrite, and eliminate the influence of serpentine on the flotation of pyrite.  . Relation between total potential energy of interaction between serpentine and pyrite particles in aqueous media and particle spacing.

Effect of Surface Potential Regulation on the Flotation Separation of Pyrite and Serpentine
The effect of surface potential regulation on the flotation separation of pyrite and serpentine was investigated by a flotation test of artificial mixed ore, as shown in Figure 8. In Figure 8a, we can see that the single mineral of pyrite had good floatability at pH < 10, and the flotation recovery was able to reach more than 90%. The addition of serpentine significantly reduced the flotation recovery of pyrite. After removing magnesium from the serpentine, the flotation recovery of pyrite in the artificial mixed ore increased, and the floatability of pyrite was restored. After removing magnesium from the serpentine and adding Mg 2+ ions, the flotation recovery of pyrite in the artificial mixed ore decreased rapidly at pH > 8, and the floatability of pyrite was affected by serpentine again. As the surface charge properties of serpentine can be significantly reduced by chain polyphosphate, the effects of sodium pyrophosphate, sodium tripolyphosphate, and sodium hexametaphosphate on the flotation separation of serpentine and pyrite were investigated, as shown in Figure 8b. The effect of chain polyphosphates on the flotation of artificial mixed ores was assessed under different pH values. It can be seen from the graph that all three kinds of polyphosphates were able to improve the flotation recovery of pyrite in the mixed ore, restore the floatability of pyrite, and eliminate the influence of serpentine on the flotation of pyrite.

Minerals
10 of 14 Figure 7. Relation between total potential energy of interaction between serpentine and pyrite particles in aqueous media and particle spacing.

Effect of Surface Potential Regulation on the Flotation Separation of Pyrite and Serpentine
The effect of surface potential regulation on the flotation separation of pyrite and serpentine was investigated by a flotation test of artificial mixed ore, as shown in Figure 8. In Figure 8a, we can see that the single mineral of pyrite had good floatability at pH < 10, and the flotation recovery was able to reach more than 90%. The addition of serpentine significantly reduced the flotation recovery of pyrite. After removing magnesium from the serpentine, the flotation recovery of pyrite in the artificial mixed ore increased, and the floatability of pyrite was restored. After removing magnesium from the serpentine and adding Mg 2+ ions, the flotation recovery of pyrite in the artificial mixed ore decreased rapidly at pH > 8, and the floatability of pyrite was affected by serpentine again. As the surface charge properties of serpentine can be significantly reduced by chain polyphosphate, the effects of sodium pyrophosphate, sodium tripolyphosphate, and sodium hexametaphosphate on the flotation separation of serpentine and pyrite were investigated, as shown in Figure 8b. The effect of chain polyphosphates on the flotation of artificial mixed ores was assessed under different pH values. It can be seen from the graph that all three kinds of polyphosphates were able to improve the flotation recovery of pyrite in the mixed ore, restore the floatability of pyrite, and eliminate the influence of serpentine on the flotation of pyrite.

Mechanism Model of Surface Potential Regulation on the Serpentine Surface
Serpentine belongs to the TO-type layered silicate minerals, and its structural unit layer is composed of magnesia-oxygen octahedrons and silica-oxygen tetrahedrons. Serpentine dissociation exposes the magnesia-oxygen octahedral layer in the crystal structure, OH-preferential dissociation occurs in the serpentine-magnesia-oxygen octahedral in aqueous solution, and residual magnesium is exposed on the surface, which makes the serpentine surface positively charged. Mg 2+ ions are one of the positioning ions in the serpentine double layer. Removal of Mg 2+ ions from the serpentine surface can reduce the positive charge density of positioning ions in the double layer, thus reducing the surface potential of serpentine. Figure 9 is a schematic diagram showing that the Mg 2+ ion dissolution and phosphate complexation reduce the surface potential of serpentine. The removal of magnesium ions from the serpentine surface reduces the positive charge density of positioning ions in the serpentine double layer, thereby reducing the surface charge properties of serpentine. The reverse adsorption of magnesium ions to the serpentine surface in solution will re-increase the positioning in the double layer. The positive charge density of ions then increases the surface charge properties of the serpentine. Therefore, the surface potential of serpentine can be regulated by the Mg 2+ ion dissolution on the serpentine surface. On the one hand, magnesium ions should be transferred from the serpentine surface to solution; on the other hand, the dissolved Mg 2+ ions can be prevented from re-entering the double layer of the serpentine solid-liquid interface. Sodium hexametaphosphate is easily hydrolyzed in aqueous solution to form polyphosphate anions. In the pulp, these anions are adsorbed on the serpentine surface by chemical interaction with the metal particles onto the serpentine surface, which causes a negative charge on the serpentine surface.

of 14
Serpentine belongs to the TO-type layered silicate minerals, and its structural unit layer is composed of magnesia-oxygen octahedrons and silica-oxygen tetrahedrons. Serpentine dissociation exposes the magnesia-oxygen octahedral layer in the crystal structure, OH-preferential dissociation occurs in the serpentine-magnesia-oxygen octahedral in aqueous solution, and residual magnesium is exposed on the surface, which makes the serpentine surface positively charged. Mg 2+ ions are one of the positioning ions in the serpentine double layer. Removal of Mg 2+ ions from the serpentine surface can reduce the positive charge density of positioning ions in the double layer, thus reducing the surface potential of serpentine. Figure 9 is a schematic diagram showing that the Mg 2+ ion dissolution and phosphate complexation reduce the surface potential of serpentine. The removal of magnesium ions from the serpentine surface reduces the positive charge density of positioning ions in the serpentine double layer, thereby reducing the surface charge properties of serpentine. The reverse adsorption of magnesium ions to the serpentine surface in solution will re-increase the positioning in the double layer. The positive charge density of ions then increases the surface charge properties of the serpentine. Therefore, the surface potential of serpentine can be regulated by the Mg 2+ ion dissolution on the serpentine surface. On the one hand, magnesium ions should be transferred from the serpentine surface to solution; on the other hand, the dissolved Mg 2+ ions can be prevented from re-entering the double layer of the serpentine solid-liquid interface. Sodium hexametaphosphate is easily hydrolyzed in aqueous solution to form polyphosphate anions. In the pulp, these anions are adsorbed on the serpentine surface by chemical interaction with the metal particles onto the serpentine surface, which causes a negative charge on the serpentine surface. Figure 9. Schematic diagram of the surface potential regulation of serpentine by Mg 2+ ion dissolution and phosphate complexation. (In the middle is the original serpentine surface double layer. In order to reduce the surface potential of serpentine, on the left side, the surface of serpentine is complexed by adding phosphates, and on the right side, the surface of serpentine is treated by pickling.).

Conclusions
In summary, we investigated the effect of the surface charge regulation of serpentine on mineral flotation, and we studied the mechanism of Mg 2+ ion dissolution and complexation from the serpentine surface to liquid and its effect on its surface charge properties. The following conclusions were drawn: (1) Strengthened Mg 2+ ion dissolution from the serpentine surface to the liquid phase can reduce the positive charge density of Mg 2+ positioning ions in the double layer. This reduces the surface potential of serpentine, thereby eliminating heterogeneous condensation between serpentine Figure 9. Schematic diagram of the surface potential regulation of serpentine by Mg 2+ ion dissolution and phosphate complexation. (In the middle is the original serpentine surface double layer. In order to reduce the surface potential of serpentine, on the left side, the surface of serpentine is complexed by adding phosphates, and on the right side, the surface of serpentine is treated by pickling.).

Conclusions
In summary, we investigated the effect of the surface charge regulation of serpentine on mineral flotation, and we studied the mechanism of Mg 2+ ion dissolution and complexation from the serpentine surface to liquid and its effect on its surface charge properties. The following conclusions were drawn: (1) Strengthened Mg 2+ ion dissolution from the serpentine surface to the liquid phase can reduce the positive charge density of Mg 2+ positioning ions in the double layer. This reduces the surface potential of serpentine, thereby eliminating heterogeneous condensation between serpentine and pyrite, and achieving good dispersion between mineral particles. (2) Chain-like polyphosphates can reduce the surface potential of serpentine. The mechanism mainly includes three aspects: They can enhance the dissolution of Mg 2+ ions from the serpentine surface to the liquid phase and reduce the surface potential of serpentine; they can interact with Mg 2+ in the liquid phase to form a stable soluble complex, thus maintaining the surface of serpentine; and the surface potential of serpentine is further reduced by adsorption onto the surface of serpentine. (3) Under natural conditions, the interfacial interaction between serpentine and pyrite in aqueous solution shows mutual attraction and heterogeneous coagulation between mineral particles. After adjusting the surface charge properties of serpentine, the interfacial interaction between serpentine and pyrite particles showed mutual exclusion, and mineral particles were dispersed in the solution system.

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