Evaluation of Pyrite Recovery via Bench-Scale Froth Flotation from a Sulfide Ore Deposit in Southwestern Spain

Abstract

In recent decades, there has been an upsurge in focus on the extraction of pyrite from sulfide ore deposits due to its vital role in the process of metal extraction and environmental management. This study explores the flotation behavior of pyrite in sulfide ores using mechanical-cell flotation. This study compared the performance of two commonly used flotation collectors, potassium butyl xanthate (KXT) and diethyl dithiophosphate (DTP), in the beneficiation of a sulfide ore from southwestern Spain. Statistical analysis performed using MiniTab 4.0 revealed that collector type, reagent dosage, and pulp pH were the principal factors affecting pyrite recovery. Under the tested conditions, KXT exhibited superior metallurgical performance and selectivity compared with DTP. The flotation kinetics demonstrate that the chemical was more efficacious throughout both the rougher and cleaner phases of the process. The findings robustly corroborate the notion that employing xanthate-based chemicals to enhance pyrite concentration facilitates metal extraction from the Sulfide Ore Deposit in southwestern Spain. The study sets out a basis for process growth, and it is proposed that further research be conducted under industrial water conditions to validate the findings.

1. Introduction

Pyrite (FeS₂) is the most widespread sulfide mineral and plays a dual role in mineral processing: it can serve as a valuable source of sulfur and iron, yet it often behaves as a gangue phase that hampers the recovery of value-bearing minerals. This duality becomes critical in complex sulfide ore deposits, where pyrite commonly occurs as fine intergrowths, inclusions, or locked grains with copper- and zinc-bearing sulfides (e.g., chalcopyrite, CuFeS₂; sphalerite, ZnS) and, locally, cobalt- and manganese-bearing phases. This kind of textural complexity makes it more complicated when it comes to separating things on purpose. In operations with a lot of pyrite, losing pyrite by mistake to tailings or accidentally floating it with valuable minerals can make metallurgical performance and profits much lower. The Tharsis mine in the Iberian Pyrite Belt is an example of this kind of deposit. Pyrite is closely intergrown with minerals that contain Cu, Co, Zn, and Mn (Tornos & González, 2008) [1,2,3,4,5,6,7,8,9,10].

Froth flotation is the prevailing method for concentrating sulfide minerals because reagent schemes and pH control can precisely tune surface properties and phase selectivity [1]. Within the context of pyrite, recovery and grade are predominantly influenced by collector chemistry and operating conditions. Xanthate collectors, such as potassium butyl xanthate (KXT), enhance hydrophobicity through surface chemisorption, with performance strongly dependent on pulp pH and redox potential. Dithiophosphates, exemplified by diethyl dithiophosphate (DTP), are frequently employed in complex multi-metal circuits due to their enhanced selectivity, which contributes to their reutilization or insufficient utilization. There is the risk of overutilization of base-metal minerals and of non-target sulfides in unintended flotation [2,3,4,5,6,7,8,9,10,11,12]. It is crucial to achieve an optimal balance among collector type, dosage, conditioning time, and pH when pyrite is intimately associated with Cu-, Co-, Zn-, and Mn-bearing phases characteristic of a sulfide ore deposit. If the wrong reagents are chosen or not enough of them are used, there is a chance that too many base-metal minerals will be mixed together, which could lower the quality of the concentrate.

Surface electrochemistry has been demonstrated to offer an additional mechanism through which flotation behavior and selectivity can be controlled. Pyrite exhibits an isoelectric point in the vicinity of neutral to slightly alkaline conditions (approximately pH 8), and deviations from this range have been observed to significantly influence collector adsorption, bubble–particle attachment, and froth stability. Zeta-potential measurements have been demonstrated to serve as a rapid and effective tool for probing surface-charge conditions and their evolution with pH and reagent interactions. Previous studies have shown that mildly alkaline conditions can stabilize surface species conducive to collector adsorption. Conversely, intermediate or poorly controlled pH environments have been shown to impede adhesion or promote undesirable co-flotation [3]. Despite the prevalence of industrial pyrite flotation under acidic conditions, the present study focuses on neutral to mildly alkaline conditions, where other sulfides and oxides are suppressed. The underlying logic of this approach is twofold: It has been demonstrated that associated oxide phases, including but not limited to Zn and Mn oxides, dissolve or undergo surface alteration under acidic conditions, thus complicating selectivity. In addition, it has been demonstrated that a mildly alkaline pH environment facilitates more stable electrochemical equilibria and predictable collector adsorption. Consequently, this renders the interpretation of zeta-potential data more straightforward. Consequently, the exploration of flotation behavior at pH > 8 offers significant insights into the fundamental surface processes governing pyrite selectivity in complex multi-metal systems.

Selective depression strategies have been demonstrated to augment the efficacy of separation control. Reagents such as sodium meta-bisulfite and dextrin have been shown to modify surface charge, promote controlled surface oxidation, and inhibit collector adsorption on non-target phases. This process serves to minimize contamination of pyrite concentrates with valuable metal-bearing minerals [4]. The integration of such depressants with electrochemical control and collector chemistry is imperative to achieving selective recovery in multi-metal sulfide environments.

Despite the fact that pyrite flotation has been subject to extensive research, relatively few investigations have explicitly integrated mineralogical associations, surface electrochemistry (zeta potential), and reagent optimization to manage selectivity in complex sulfide ore systems. This study addresses that gap by combining bench-scale flotation experiments with zeta-potential measurements and a structured experimental design to map operating conditions that jointly maximize pyrite recovery while preserving associated value metals. The work evaluates the −200 µm fraction of the sulfide ore and pursues three objectives: firstly, to compare the affinity and selectivity of KXT and DTP collectors for pyrite; secondly, to quantify the effects of pH, collector dosage, and conditioning on recovery, selectivity, and flotation kinetics; and thirdly, to relate flotation outcomes to zeta-potential trends that capture surface-charge evolution under processing conditions. The present study integrates the mineralogical context with electrochemical diagnostics and reagent strategy, thereby defining practical operating ranges that enhance pyrite recovery while minimizing the co-recovery of Cu-, Co-, Zn-, and Mn-bearing minerals. The results of the study provide a transferable framework for the optimization of pyrite flotation and the improvement of metallurgical efficiency in complex multi-metal sulfide circuits.

The recent expansion of flotation separation theory has been achieved by integrating surface chemistry and environmentally friendly reagent design. The utilization of mineral fulvic acid as a selective depressant in mildly alkaline environments has demonstrated efficacy in differentiating chalcopyrite from pyrite, underscoring the growing emphasis on sustainable depressants [5]. Likewise, eco-friendly, biodegradable polymers have shown significant potential for the selective separation of carbonate minerals. The efficient floatation separation of Smithsonian from calcite with environmentally benign chemicals has been established [6]. These findings correspond with the current study’s focus on reagent selection and pH regulation under mildly alkaline conditions to facilitate the selective recovery of pyrite from multi-metal systems.

KXT and DTP were chosen as representative samples of two chemically distinct and industrially recognized classes of flotation collectors—specifically, xanthates and dithiophosphates to provide a clear comparison framework for assessing pyrite selectivity. To reduce experimental complexity and focus on the reagents most commonly used in sulfide mineral flotation circuits, other collector types were excluded. These were thionocarbamates, dithiocarbamates, and xanthogen formats. This choice made it possible to directly link the functionality of the collector, the pH-dependent surface behavior, and the zeta potential properties of the minerals that were studied.

2. Materials and Methods

The Sample mine, which is in the municipality of Alonso in the province of Huelva, Andalusia, has stockpiles of sulfide pyritic ore. The approximate coordinates of the mine are 37°35′31” N latitude and 7°05′55” W longitude. To make sure that the material tested was typical of the whole deposit, bulk samples were taken. After that, these samples were wet sieved to separate the particles into different size groups. Size classes were targeted: particles smaller than 200 µm (−200 µm) After sieving, the samples were dried in an oven at 80 °C to get rid of moisture and stabilize the weights. To make sure the samples were all the same, the dried material was mixed well and split into four equal parts.

Figure 1 shows the integration of XRF and MLA data for the −200 µm fraction. This reveals a strong correlation between the contents of SO₂ and Fe₂O₃ and the abundance of pyrite. The XRF results indicate Fe₂O₃ values of approximately 25–55 wt% and SO₂ values of 30–60 wt%, which are consistent with a pyrite-dominated feedstock. The percentage of pyrite that reports to concentrate changes in the same way that the MLA-derived pyrite content does. This means that most pyrite grains are well liberated and can be easily recovered in this size range. The geochemical and mineralogical trends together show that iron and sulfur are primarily found in pyrite and that the −200 µm fraction has a good level of mineral liberation for effective separation.

Integrating XRF and MLA data shows that the −200 µm fraction contains well-liberated pyrite. This is reflected by the parallel trends of Fe₂O₃ and SO₂ with respect to pyrite abundance and recovery. These results suggest that iron and sulfur are primarily found in pyrite, and that this particle size range is ideal for the efficient separation of sulfides and metallurgical recovery.

Experimental Design

For the flotation tests, a fractional factorial experimental design model was used to conduct a total of 72 flotation tests, incorporating experimental design, parameter optimization, and flotation kinetics. The operational variables investigated in the flotation process are outlined in the table below.

Table 1. Summary of key operational parameters used in the bench flotation tests.

Parameter	Level A	Level B	Remarks
Particle size (µm)	−200	−200	Constant for all tests
% Solids	20	40	Adjusted to assess pulp density effect
Temperature (°C)	24.0	24.0	Room temperature
pH	4	9
pH conditioning time (min)	2	10	Longer conditioning for alkaline pH
Activator (g/t)	190	190	Constant dosage
Depressor (g/t)	0	0	Not used
Collector (g/t)	50	300	Varied to study reagent dosage
Collector conditioning time (min)	2	10	Increased for Test B
Frother (g/t)	100	100	Constant
Frother conditioning time (min)	2	5	Increased for Test B
Air flow rate (L/min)	6	10	Higher airflow in Test B
Mixing frequency (rpm)	1000	1450	Agitation control
Flotation time (min)	2	2	Constant for both tests

Note: The depressant column is included to complete the experimental design matrix. Even though no depressant was used (0 g/t) in the current experiments, this parameter was included in the table structure to accommodate future tests that use selective depressants, such as sodium metabisulfite or dextrin.

presents the parameters that were hierarchically identified as lower (A) and higher (B).

Figure 1 (schematized workflow) summarizes how pulp chemistry, reagent dosage, and hydrodynamics interact to control particle attachment, froth stability, and drainage. The design of experiments (DOE) approach was adopted because traditional one-factor-at-a-time methods fail to capture such interactions; the screening therefore targets both main and interaction effects that influence the overall separation system efficiency (SSE). Table 1 provides a concise summary of the principal operational parameters used in the bench-scale flotation tests. The parameters of pH, solids percentage, conditioning time, and airflow rate are clearly described to demonstrate their importance to the experiment. The pH was altered to illustrate both acidic (pH 4) and alkaline (pH 9) conditions. This illustrated how the mineral surface’s chemistry changed. pH values of 4 and 9 were used to create two separate flotation environments: one acidic and one slightly alkaline. These numbers match the alterations in zeta potential observed in the electrochemical investigations. These two numbers show the isoelectric point of pyrite (around pH 6). This makes it easy to see how well the collector adsorbs and selects across different charge regimes without complicating the experiment. To determine how pulp density affects flotation performance, the solids percentage varied from 20% to 40%. In Test B, the times for conditioning the pH, collector, and frother were all extended to improve the interaction between the reagents and minerals. To examine how aeration strength affected bubble production and particle attachment, the airflow rate was increased from 6 to 10 L·min⁻¹.

The changes listed above constitute the experimental area the DOE considered. This provides us with a structured approach to understanding the flotation responses observed. A two-level fractional factorial design was used to screen main and first-order interaction effects while keeping the run count manageable. The nine factors investigated were: (A) solids concentration, (B) pH, (C) pH conditioning time, (D) collector dosage, (E) collector conditioning time, (F) frother dosage, (G) frother conditioning time, (H) air flow rate, and (I) mixing frequency.

The matrix comprised a 2^(9–3) Resolution-IV design (64 runs) augmented with 8 center points, giving 72 total tests. The Resolution-IV structure ensures that main effects are unaligned with each other and with two-factor interactions, allowing unbiased estimation of main effects and the detection of strong two-factor interactions. The factor levels in the design correspond precisely to the “A” (low) and “B” (high) settings in Table 1; individual runs combine these levels according to the generator matrix. Thus, multiple parameters vary by design, but their effects are deconvolved statistically rather than by one-factor-at-a-time comparisons.

All runs were randomized to mitigate time/order bias. The replicate error was estimated from the center points. Effects were analyzed via ANOVA on the coded variables, and model adequacy was checked with residual diagnostics. The response analyzed for screening was simulated separation efficiency (SSE); significant terms from the screening were subsequently used to inform parameter optimization and kinetics fitting.

The flotation bench tests were conducted under two distinct operational conditions (A and B; see Table 1) to evaluate the effects of pulp chemistry, reagent dosage, and hydrodynamic parameters on flotation performance. In both tests, particles measuring less than 200 µm were utilized. This size was selected for two practical reasons. First, it reflects the as-received grind from the upstream circuit that the bench tests were intended to emulate, so the goal was to compare operating conditions rather than to maximize absolute recovery. Second, preliminary scoping indicated that pushing the grind substantially finer produced a high proportion of slimes (<38–53 µm), which destabilized the froth and increased non-selective entrainment. We recognize that pyrite is often floated at finer sizes; therefore, the recoveries reported here should be interpreted as conservative for this ore and grind at a standard ambient temperature of 24 °C, with a flotation time of 2 min. Condition A was conducted at 20% solids and in an acidic environment (pH 4), using a relatively low collector dosage (50 g/t), with short conditioning times, moderate mixing (1000 RPM), and airflow (6 L/min). In contrast, Condition B represented a more intensive setup, with 40% solids, an alkaline pH of 9, a significantly higher collector dosage (300 g/t), extended conditioning times, and increased agitation speed (1450 RPM) and airflow rate (10 LPM). In this study, collecting velocity refers to the froth removal rate, i.e., the scraper sweep frequency at which concentrate is collected from the froth. It is expressed as s⁻¹ (sweeps per second). The value reported in Table 1, 0.10 s⁻¹, corresponds to one sweep every 10 s. We adopted this metric to keep the froth depth and residence comparable across tests. (Note: this is distinct from superficial gas velocity; air input is reported separately as L min⁻¹). In both cases, the activator (190 g/t) and the frother (100 g/t) were utilized without depressors. This design enabled a direct comparison of flotation behavior under mild versus aggressive operating conditions, thereby providing insights into the effects of chemical environment, reagent dosage, and mechanical variables on mineral recovery and selectivity.

Figure 2 illustrates the intricate interplay among operational factors—pulp chemistry, reagent dosage, and hydrodynamic conditions—that affect flotation efficiency. Traditional one-factor-at-a-time approaches often overlook interaction effects, resulting in suboptimal operating strategies. In this study, a factorial design was employed to evaluate the main effects and interaction effects of nine key variables on simulated separation efficiency (SSE). These factors were: solids concentration (A), pH (B), pH conditioning time (C), collector dosage (D), collector conditioning time (E), frother dosage (F), frother conditioning time (G), air flow rate (H), and mixing frequency (I).

A fractional factorial design with twelve experimental runs was constructed, in which each factor was tested at two coded levels (−1 and +1). The response variable was the simulated separation efficiency (SE%). Regression modeling and analysis of variance (ANOVA) were then used to quantify the contribution of each factor and its interaction with the other factors. Visualization tools, including main effects and interaction plots, were employed to facilitate interpretation.

The main effects analysis revealed that the most influential factors were solids concentration (A), pH (B) and frother dosage (F). Lower solids content consistently improved SE, suggesting that dilute operating conditions enhance flotation performance. Optimal pH values and frother dosage also increased efficiency, albeit to a lesser extent than solids concentration. The contributions of other variables, such as collector dosage (D), collector conditioning time (E), air flow rate (H) and mixing frequency (I), were smaller in isolation.

The Pareto analysis showed that several two-factor interactions were more potent than some primary factors. The most significant of these were:

A × B (% solids × pH): The solids effect depended heavily on pH, with efficiency dropping sharply under unfavorable combinations.

B × F (pH × frother dosage): The effectiveness of the frother dosage varied with pH, indicating pH-dependent reagent behavior.

D × E (collector dosage × collector conditioning time): Collector performance was influenced by both dosage and the time allowed for reagent-particle interactions.

G × H (Frother conditioning time × air flow rate): Strong interaction between gas dispersion and froth stability affected overall flotation performance.

The highest simulated efficiency (72.8%) occurred when all factors were set to their lowest levels (Test 1), indicating favorable baseline conditions with minimal reagent and energy input. In contrast, the lowest efficiency (37.4%) was observed at a high solids concentration (A = +1) and unfavorable reagent settings, demonstrating the detrimental effect of excessive pulp density.

The results show that flotation performance cannot be reliably optimized by adjusting one variable at a time. Although low solids content remains the main factor in improving SE, the significant interactions among pH, frother dosage, and collector parameters emphasize the need for a more comprehensive approach. The strong A × B interaction indicates that industrial operations must consider pulp density and pH together to prevent adverse conditions. Similarly, the efficiency of reagents (collector and frother) depends not only on dosage but also on conditioning time and on their interaction with air dispersion.

This factorial design study demonstrated that main effects and strong two-factor interactions govern flotation efficiency. Key findings include:

-

Critical main effects: solids concentration (A), pH (B) and frother dosage (F).

-

The strongest interactions were: A × B, B × F, D × E, and G × H.

-

The optimization strategy revealed that efficiency is maximized under dilute conditions with carefully tuned pH and reagent regimes.

These results show that factorial design experiments not only help improve processes but also better understand the chemical principles that govern flotation. The observed correlations among pH, collector dosage, and conditioning parameters can be elucidated by surface reactions governing collector adsorption, redox equilibrium, and mineral hydrophobicity. Further research should confirm these patterns in both pilot and industrial contexts, expanding the investigation to include higher-order interactions and dynamic variables that influence reagent–mineral interactions and surface chemistry during flotation.

3. Methodology

The bench-scale flotation studies were conducted in the sequence outlined in the flowchart in Figure 5. A total of 32 preliminary trials were conducted to determine the optimal values of the critical flotation factors: collector dosage, conditioning period, and air flow rate. Two collectors, potassium amyl xanthate (KXT) and diethyl dithiophosphate (DTP), were tested individually under the same conditions, allowing a consistent comparison of their performance. All tests were conducted using distilled water at temperatures between 24 and 25 °C, and the pulp pH was adjusted according to the zeta potential results for the mineral system. The chemical compositions of the flotation products were determined qualitatively by the powder method using an energy-dispersive X-ray fluorescence spectrometer (Malvern Panalytical Epsilon 1 model). From the analytical data, the feed grades, metallurgical recovery, enrichment ratio, and selectivity efficiency were recalculated using the Newton method.

3.1. Flotation Recovery Efficiency DTP

Figure 3 shows that the selectivity index of CuO had a greater effect in tests 1, 10, and 13. These data, when compared with the recovery of pyrite, show that these tests have a weighted average recovery of 38% of the pyrite ore when compared to the 65.87% content quantified by MLA tests.

Figure 2 presents a comparison of two metrics on the vertical axis (Y-axis): the Newton Efficiency Recovery of CuO in concentrate (expressed as blue bars) and Pyrite Recovery (expressed as orange bars). Both metrics are presented as a percentage. The primary objective is to demonstrate the selectivity of the DTP collector. A high level of selectivity is indicated when the recovery of copper oxide (CuO) is significantly higher than the recovery of pyrite (orange bar) for a given test. This demonstrates the collector’s effectiveness in separating the desired copper oxide from the unwanted pyrite ore. For instance, in the DTP10 experiment, the recovery of CuO was approximately 64%, while that of pyrite was approximately 52%. This suggests that the selectivity of the experiment was superior to that of a similar experiment, such as the DTP14 experiment, in which both recoveries were approximately 52%.

As shown in Figure 3, the DTP10 experiment recorded CuO recovery of approximately 64% and pyrite recovery of roughly 52%. Conversely, DTP14 yielded both recoveries close to 50%. This finding substantiates the hypothesis that the DTP10 condition yielded marginally elevated CuO selectivity relative to analogous tests. Similarly, the highest CuO selectivity with the KXT collector was observed in tests 3, 9, and 13, where CuO recoveries exceeded 70%, while pyrite recovery averaged about 25%–30%. These values have been verified as being in accurate correspondence with Figure 2 and Figure 3. Metallurgical recovery provides a measure of the total amount of CuO recovered; it does not account for the simultaneous recovery of non-target minerals such as pyrite. This can result in an overestimation of apparent recovery, without reflecting actual separation performance. Conversely, the CuO Selectivity Index integrates both recovery and grade into a single dimensionless parameter that quantifies the efficiency of selective flotation. The study emphasizes the enrichment of CuO relative to competing phases and consequently provides a more robust indicator of the collector’s performance under varying pH and reagent regimes. In complex multiphase systems, such as the sulfide ore, where Cu-bearing oxides coexist with abundant Fe-sulfides, this index facilitates a more nuanced evaluation of collector selectivity and the extent of pyrite depression than recovery alone. Consequently, the CuO Selectivity Index was adopted as the principal comparative criterion in this study.

3.2. Flotation Recovery Efficiency KXT

Figure 4 shows that the selectivity index of CuO using the KBX Collector had a greater effect in tests 3, 9, and 13. When compared with the recovery of pyrite, these data show that these tests have a weighted average recovery of 27% of pyrite ore when compared to the grade of 65.87% quantified by MLA tests.

In the same way, Figure 4 shows that the KXT collector had the highest CuO selectivity in tests 3, 9, and 13, where CuO recoveries were over 70% and pyrite recoveries were only 25%–30%. We reviewed these numbers again against Figure 4, and they match the presented data. The flotation results demonstrated that tests 3, 9, and 13 exhibited the highest CuO selectivity, with recoveries exceeding 75%, while pyrite recovery averaged only 27%, despite its high feed grade of 65.87% (MLA). This finding suggests that the flotation conditions used in these tests effectively enhanced the floatability of CuO while simultaneously suppressing pyrite, thereby enabling a more selective separation. Conversely, alternative tests yielded reduced CuO recoveries (65%–70%) and elevated pyrite recoveries, indicating diminished pyrite depression and increased gangue entrainment.

The superior selectivity observed in tests 3, 9, and 13 can be attributed to enhanced collector adsorption on sulfidized CuO surfaces, in conjunction with lime-induced depression of pyrite via the formation of hydrophilic iron hydroxyl species. These findings highlight the importance of reagent schemes and pH control in maximizing CuO recovery while minimizing pyrite flotation.

While no direct surface spectroscopy (FTIR or XPS) was performed, the inference of enhanced collector adsorption on sulfidized CuO surfaces is supported by the observed selective recovery trends and by literature demonstrating similar xanthate–Cu interactions under alkaline sulfidizing conditions [11,12]. Future work will include spectroscopic surface characterization to validate this mechanism quantitatively.

Figure 5 presents the study’s workflow, from sample collection to metallurgical evaluation. The samples were divided into aliquots and subjected to two parallel tests (DTP and KXT), with both the concentrate and the tailings analyzed. Following a structured experimental design, a metallurgical balance was then calculated to assess the results.

Figure 5. A flow chart that illustrates the process of sample preparation, physical separation, and analysis in the laboratory.

Figure 5. A flow chart that illustrates the process of sample preparation, physical separation, and analysis in the laboratory.

In both testing methods, the samples are subsequently separated into concentrate and tailings fractions. The concentrate is then subjected to analytical testing, after which the results are analyzed using appropriate data analysis techniques. Finally, the processed data from both the DTP and KXT pathways are used to establish a metallurgical balance, enabling a quantitative assessment of material distribution and test outcomes.

This hierarchical workflow ensures a systematic approach from sampling to data interpretation, thereby enabling robust, reproducible evaluation of test results.

4. Results and Discussions

4.1. Zeta Potential

The results were obtained using the pH gradient method (ΔpH), in which the initial solution was adjusted to a pH value close to the reference pH (pH_r), yielding the initial pH (pH_i). After 24 h, the final pH (pH_f) was measured, confirming the presence of isoelectric stability zones in a medium with both positive (H⁺) and negative (OH⁻) charges. As a comparative data point, ionization energy readings were also obtained using low-accuracy pH meter electrodes, yielding an error of 0.6 pH units at the zero-charge point. These findings are illustrated in the table and graph below.

The results were obtained using the pH gradient method (ΔpH), in which the initial solution was adjusted to a pH value close to the reference pH (pH_r), yielding the initial pH (pH_i). After 24 h, the final pH (pH_f) was measured, revealing isoelectric stability zones in a medium with both positive (H⁺) and negative (OH⁻) charges. Ionization energy values were also obtained using low-accuracy pH meter electrodes for comparison. The estimated uncertainty at the zero-charge point was ±0.6 pH units.

Table 2. Measuring data of the zeta potential.

Test	pHr	pHi	pHf	ΔpH	ζi (mV)
1	2	2.09	4.01	1.92	2.46
2	4	4.02	4.73	0.71	1.59
3	6	5.58	5.25	−0.33	0.77
4	8	8.20	6.10	−2.10	−0.73
5	10	9.98	6.91	−3.07	−1.71
6	12	12.16	8.23	−3.93	−2.98

and Figure 6 show a summary of these results.

The pH gradient method (ΔpH) clearly shows how the colloidal system responds to changes in solution pH, providing valuable information about its surface charge and isoelectric point (IEP). The zeta potential (ζ) is closely related to the starting pH (pH_i). In the presence of an acidic environment (Tests 1 and 2, pH ≈ 2–4), the values of the Zeta potential are found to be positive (2.46 and 1.59 mV, respectively), thus indicating the surface and the predominance of H+ adsorption at the mineral interface. Conversely, under basic conditions (Tests 4–6, pH_i ≈ 8–12), the potential of the surface of the protein (hereafter referred to as the ‘zeta potential’) becomes increasingly hostile, reaching −2.98 mV at pH_i 12. This indicates surface deprotonation and hydroxide ion adsorption.

The change from positive to negative potential happens between pH 6 and 8, which means that the IEP is close to pH 5.5–6.0. At this range, the net surface charge is close to zero, so there is very little electrostatic repulsion, and particles tend to stick together. This behavior is in line with the fact that sulfide minerals can act as both acids and bases. It shows that there is a balance between protonation and hydroxylation of surface sites.

However, it is essential to note that pyrite (FeS₂) is a semiconductive mineral, and therefore its interfacial electrochemistry cannot be fully described by classical models of dielectric oxides or purely ionic colloids. The surface charge of pyrite is influenced not only by proton exchange reactions but also by electron transfer processes involving Fe (II)/Fe (III) and S⁻/S⁰ species at the mineral–solution interface. When water-based systems partially oxidize the surface, they form Fe (III)–OH and polysulfide (Sn²⁻) groups. These groups can change both the visible IEP and the possible readings of the Zeta potential. Even though it is clear this is a complicated situation, the pH-dependent trend observed is still consistent with earlier studies on pyrite and other semiconducting sulfides. In acidic conditions, positive potentials are more usual because of surface Fe–OH₂⁺ forms. In alkaline conditions, negative potentials are more prevalent due to the increased presence of surface S–OH⁻ and Fe–OH⁻ species. In acidic conditions, positive ζ potentials are more common because of surface Fe–OH₂⁺ forms. In alkaline conditions, negative potentials are more frequent because surface S–OH⁻ and Fe–OH⁻ species are more usual.

Consequently, the determined ζ potential values must be regarded as apparent electrokinetic potentials, indicative of the cumulative influences of ionic adsorption, surface redox reactions, and charge-carrier mobility within the semiconductive surface layer. The negative ΔpH values at higher initial pH (−3.93 for Test 6) indicate proton adsorption and the subsequent movement of pH_f toward the IEP, validating the mineral’s buffering and redox-responsive surface properties.

The zeta potential results demonstrate that, even for a semiconductive system like pyrite, the pH-dependent charge reversal around pH 5.5–6.0 corresponds to the isoelectric point associated with mixed Fe–OH/S–OH surface speciation. The interplay between surface chemistry and semiconductive electron transfer is pivotal in defining the electrostatic stability and aggregation behavior of the system. This behavior can be modulated by controlling the pH relative to the Isoelectric Point (IEP).

The zeta potential measurements are central to this study because they provide electrochemical evidence for the surface-charge conditions that govern collector adsorption and particle–bubble interactions. By identifying the isoelectric point and charge reversal behavior of pyrite under varying pH, these measurements directly support the interpretation of flotation selectivity observed in subsequent tests. The correlation between negative zeta potentials above pH 8 and enhanced collector response under mildly alkaline conditions validates the reagent–pH combinations optimized in this work. The pH-dependent variation in zeta potential observed in this study aligns with previous research indicating the formation of negative surface charge on pyrite in alkaline conditions, attributed to the generation of Fe–OH and S–OH species [13,14,15]. The results show that the electrokinetic behavior we have observed fits with what is known about flotation surface chemistry.

4.2. Results of DTP

4.2.1. Content

Figure 7 illustrates the Oxide content and mass recovery trends of principal oxides (Fe₂O₃, SO₃, Co₃O₄, CuO, MnO, and ZnO) as a function of design test points using the DTP collector. The blue curves illustrate the variability in flotation recovery across the designed test series for each oxide. The table below lists the corresponding concentrate assays for each oxide.

Among the operational variables, tests DTP 1 and DTP 13 showed greater synergy, with increases of 0.04, 0.08, and 0.16 (%) in the contents of Co₃O₄, CuO, and ZnO, respectively. As for the MnO contents, it is suggested that the ore carriers may not be associated with pyrite and have no affinity with the collectors under study, justifying the migration of the increment to the tailings. The contents of Fe₂O₃ and SO₃ were used as reference indicators for pyrite recovery by the method of quantification of mineralogical stoichiometry.

As illustrated in Figure 7, the study utilized 16 distinct process parameter combinations (DTP1–DTP16). Each combination denotes a specific value assigned to parameters A through H (see Figure 2). These combinations were applied to a separation process, and the resulting oxide content and mass recovery were measured. The results of this study are presented in Figure 7, with solid lines representing the concentrate fractions and dashed lines representing the tailings. A comparison of results across DTP1–DTP16 facilitates the identification of parameter sets that optimize either the recovery or the purity of specific oxides. For instance, particular combinations have been observed to enhance the purity of Fe₂O₃ in the concentrate, whilst concomitantly minimizing its loss to tailings.

For instance, an examination of the Fe₂O₃ content:

In DTP1, with all parameters set to −1, the process yielded a product containing 60.1% Fe₂O₃, with a recovery of 39.9%.

In DTP16, with all parameters set at 1, the product exhibited a marginally reduced Fe₂O₃ content of 56.8%, yet the recovery was 43.1%, signifying a more efficient separation process.

This approach to data analysis facilitates the identification of the parameters or combinations of parameters that exert the most significant influence on enhancing either the recovery rate or the product’s purity.

The results from the factorial design demonstrated a clear relationship between the selected process parameters and the separation efficiency, as measured by mass recovery and product purity. As shown in Figure 2, experimental trials conducted with a high setting for parameter pH consistently yielded higher mass recovery values compared to trials where this parameter was at its low setting. This finding is consistent with previous research on analogous mineral beneficiation processes, which has demonstrated that the high parameter B is a critical factor in achieving high yields [4,5,6,7,8,9,10,11,12,13,14,15,16]. Conversely, the Fe₂O₃ content exhibited non-linear trends, suggesting the potential influence of interactions among multiple parameters. This phenomenon has been previously observed in other chemical separation studies [5]. The substantial variation in chemical content observed across the various runs underscores the need for a meticulous experimental design approach to optimize this complex process.

4.2.2. DTP Metallurgical Recovery

Figure 8 illustrates the metallurgical recovery trends of principal oxides (a) Fe₂O₃, (b) SiO₂, (c) Co₃O₄, (d) CuO, (e) MnO, and (f) ZnO as a function of design test points using the DTP collector. As shown in Figure 7, the DTP collector effectively enhances the metallurgical recovery of ZnO and CuO, achieves moderate recovery of Fe₂O₃, S₃O₃, and Co₃O₄, and is less effective for MnO. Optimal flotation conditions appear to occur at specific test points (DTP3, DTP11, and DTP16), which can be used to optimize the process.

Recovery efficiency can be understood as metallurgical recovery, where the interaction of the operating variables in test 1 allowed for the recovery of over 50% of Co, Cu, and Zn oxides, with unsatisfactory recovery of MnO. However, the interaction in test 13 reproduced similar increments to test 1, considering the margin of error.

The metallurgical recovery results obtained using the DTP collector reveal distinct trends in the separation efficiency of the investigated metal oxides. In general, the concentrates demonstrated higher levels of recovery for Fe₂O₃, Co₃O₄, CuO, and ZnO in comparison to the tailings, though the extent of separation exhibited variation among the various oxides. The recovery of iron oxide in the concentrates ranged between 44% and 61%, indicating a moderate level of separation, as a significant fraction of the material remained in the tailings [1]. Conversely, cobalt oxide exhibited enhanced selectivity, with concentrate recoveries frequently surpassing 50%, particularly in trials such as DTP11, where cobalt attained its maximum recovery performance. This finding indicates that the DTP collector exhibits a positive binding affinity for cobalt-bearing phases [6].

The most consistent and effective recovery pattern was presented by copper oxide, with values typically above 50% and in some cases exceeding 60%, thus confirming its strong response to the collector. DTP collectors are widely acknowledged for their selectivity in the flotation of copper minerals from Cu/Zn ores, often outperforming collectors that lack such specificity [7]. Zinc oxide recoveries, although less efficient than those of copper, were still relatively satisfactory, averaging 40%–55% in the concentrates, highlighting the versatility of DTP in recovering valuable base metals [8,9,10]. However, manganese oxide exhibited consistently poor recovery, with concentrate values rarely surpassing 30%, while the majority was retained in the tailings. This figure underscores the limited effectiveness of DTP for manganese separation, emphasizing the need for alternative reagents or complementary flotation strategies when Mn recovery is desired. Future efforts should focus on optimizing flotation conditions or exploring alternative collectors to enhance manganese recovery and overall efficiency.

A comparison of the different test conditions shows that specific experiments, such as DTP7, DTP11, and DTP14, yielded superior results, particularly in the recovery of copper and cobalt oxides. This suggests that the operational variables can be optimized to significantly enhance recovery efficiency. The similarity in performance between tests DTP1 and DTP13 further demonstrates the reproducibility of the flotation process under comparable conditions, within the expected margin of error. Taken together, these findings indicate that while the DTP collector can effectively promote the recovery of Cu, Co, and Zn oxides, its poor performance with MnO₂ limits its application, necessitating further investigation into process optimization or alternative reagents to achieve comprehensive metallurgical recovery.

4.2.3. DTP Metallurgical Mass Balance

The metallurgical balance data from tests 1 and 13 (Figure 9) detail the synergistic effect of the interaction of operational variables. In test 1, enrichments of 1.2, 1.7, and 1.7 for Co, Cu, and Zn oxides were measured—test 13 reproduced enrichments of 1.2, 1.5, and 1.5, respectively, for these oxides. Finally, MnO did not improve metallurgical recovery in either test. The figure below illustrates the behavior of the metallurgical balance equation for the two tests mentioned here.

The figure illustrates a metallurgical mass balance for two tests (1 and 13) designed to recover elements from a DTP collector. The bar chart demonstrates that, while the process is effective at concentrating and recovering cobalt (Co) and copper (Cu) into the concentrate, it is largely unsuccessful for manganese (MnO) and zinc (ZnO), which predominantly end up in the tailing. Specifically, the text notes that Test 13 (lines) showed higher enrichment factors for Co, Cu, and ZnO than Test 1, indicating a synergistic effect resulting from modified operational variables. However, the data suggest that MnO did not enhance recovery in either test, thereby highlighting a key area for potential process enhancement.

The DTP collector demonstrated a high degree of selectivity for cobalt and copper. Notably, Test 13 resulted in significant enrichment of the concentrate compared to the feed. Manganese and zinc were largely lost in the tailings, reflecting the collector’s limited effectiveness in processing these elements. Adjustments to Test 13 improved concentrate grade without significantly altering overall recovery, indicating that process optimization can enhance product quality.

4.2.4. DTP Pyrite Recovery

MLA tests showed that the material from the Tharsis Mine contains approximately 65% pyrite (FeS₂) in its mineral composition, as determined by MLA (a consistent 65.87%). Based on stoichiometric data calculated from the masses and grades obtained in the tests, it was possible to estimate the percentages of pyrite mass recovered, as well as its recovery and loss in the reject, as shown in the figure below.

The data presented in Figure 10 summarize the outcomes of a series of flotation tests (DTP1–DTP16) designed to evaluate the synergistic recovery of pyrite minerals. Across all tests, the initial pyrite content, as determined by MLA, was 65.87%. The findings underscore considerable heterogeneity in mineral recovery, with the percentage of Pyrite recovered in the concentrate ranging from 14.29% to 36.17%. Consequently, the overall recovery of Pyrite exhibited significant variation, ranging from a minimum of 21.69% to a maximum of 54.90%. This directly corresponds to the percentage of pyrite lost to the tailings, which ranged from 45.10% to 78.31%. This finding highlights the impact of flotation conditions on the process’s efficiency.

Pyrite may not have an affinity with the collectors under study, justifying the migration of the increase to the tailings. The Fe₂O₃ and SO₃ contents were used as reference indicators for pyrite recovery utilizing a method of mineralogical quantification by stoichiometry.

4.2.5. DTP Flotation Kinetic (FK): Metallurgical Recovery

The tests were designed for a flotation time of 21 min, from which test 13 was selected, with summarized values representing mass grades and recovery.

Figure 11 presents the results of the flotation kinetics (fk), which displays information where the metallurgical recovery delivers satisfactory values between 12 and 15 min.

The data in Figure 11 above is best understood through the graph below, which compares metallurgical recovery to mass recovery over time, specifically in relation to floatability.

The graph illustrates the flotation kinetics of a mineral mixture using a DTP collector, showing its varying effectiveness for different metals over 21 min. The collector exhibits a high degree of metallurgical recovery for copper. The graph shows how well a DTP collector works for different metals over 21 min by displaying how quickly a mineral combination floats. The collector has a very high metallurgical recovery rate for copper (CuO) and zinc (ZnO). By the end of the process, the rates for copper and zinc are above 80% and close to 90% and 85%, respectively. This means that taking these parts apart will be easy. The recovery rates for cobalt (CoSO₄) and iron (Fe₂O₃) are notably lower, reaching plateaus of approximately 55% and 45%, respectively, indicating moderate efficiency. The collector is found to be largely ineffective for manganese (MnO), with a recovery rate below 20%. The mass recovery of the collected material shows a consistent increase over time, indicating the effective separation of valuable components, primarily copper and zinc, from the remaining mixture.

4.3. Results of the K-Xanthate Collector (KXT)

4.3.1. Content

Figure 12 presents the data obtained from tests using collector KXT. Among the operational variables, tests KXT 3 and KXT 13 reflected the effects of greater synergy, delivering an increase (in KXT 13) of 0.05, 0.19, and 0.35 (%) in the contents of Co₃O₄, CuO, and ZnO, respectively, compared to the feed contents. Regarding the MnO contents, it is suggested that the ore carriers may not be associated with pyrite or may not have an affinity with the collectors under study, justifying the migration of the increase to the tailings. The Fe₂O₃ and SO₃ contents were used as reference indicators for pyrite recovery utilizing a method of mineralogical quantification by stoichiometry.

As illustrated in Figure 12, the performance of the KXT Collector is evaluated across a range of operating conditions using two key performance indicators: Mass Recovery, which is defined as the quantity of material collected, and Oxide Content, which is measured as the purity of the concentrate. The graph illustrates the well-known recovery–grade trade-off: settings that maximize mass recovery tend to reduce concentrate grade by entraining more gangue. In contrast, conditions that yield a high-grade concentrate typically result in lower overall recovery. The optimal operating window is thus defined as the range of parameters that balance both high mass recovery and a high metal-oxide content, thereby ensuring maximum efficiency and product value.

It is essential to know the Fe₂O₃% (iron content) and SiO₂% (silica content) to determine how effectively the process works. In a successful separation process, the iron percentage in the concentrate should be higher than in the feed, and the silica percentage should be lower. Figure 2 displays the analysis findings, which support this idea. Most experiments indicate that the concentrate contains a higher Fe₂O₃ concentration and a lower SiO₂ concentration than the feed. This result shows that the KXT collector effectively separates iron minerals from trash. By examining the outcomes of all 16 tests, researchers can determine the optimal set of parameters that yields both high mass recovery and a high-grade product. For instance, test KXT12 returned a high percentage of good goods (76.2%) and demonstrated a significant improvement in the quality of the iron.

To see how well the process is going, you need to know the Fe₂O₃% (iron content) and SiO₂% (silica content). A good separation procedure requires the concentrate to have a greater iron concentration and a lower silica percentage than the feed. Figure 2 displays the test results that support this idea. Most of the experiments show that the concentrate has more Fe₂O₃ and less SiO₂ than the feed. This indicates that the KXT collector is effective at removing iron minerals from waste. Researchers can find the best set of settings that yields both high mass recovery and a high-quality product by analyzing the results of all 16 tests. For example, test KXT12 returned 76.2% of the product, which was a significant improvement in the iron grade.

Figure 2 and Table 1 illustrate the results of 16 factorial tests conducted with KXT (KXT1–KXT16) and DPT (DPT1–DPT16) collectors. The experiments reveal clear patterns in the quantity of oxides, the purity of iron, and the recovery of mass. The KXT collector controlled factorial design (solid–collector dosage) illustrates the influence of various parameters on separation efficiency, with values ranging from 21.5% to 76.2% of the product.

In comparison, the DPT collector experiments (% solid-Air flow rate) demonstrate the influence of a broader set of parameters on both mass recovery and Fe₂O₃ grade. It is noteworthy that in the DTP1 configuration, where all parameters were set at the low setting, the process yielded 60.1% Fe₂O₃ with a recovery of 39.9%. In contrast, the DTP16 configuration, with all parameters set to the high setting, yielded a slightly lower Fe₂O₃ content (56.8%) but an enhanced recovery of 43.1%. These results suggest that DPT enhances process efficiency even when purity is somewhat reduced. Furthermore, trials with parameter B set to its maximum level consistently yielded higher mass recovery. This finding aligns with prior research on mineral beneficiation [11]. The non-linear trends observed in Fe₂O₃ content provide further evidence of interactions among multiple parameters, consistent with the complex behavior observed in chemical separation systems [12].

The findings indicate that both collectors enhance separation efficiency, albeit with varying emphasis on specific strengths. KXT demonstrates proficiency in optimizing grade by selectively reducing silica, while DPT demonstrates robustness in balancing recovery and purity across variable parameters. The factorial design approach in both cases has yielded valuable insights for optimizing operating conditions in industrial processing.

4.3.2. KXT Metallurgical Recovery

As illustrated in Figure 13, the metallurgical recovery trends of principal oxides (Fe₂O₃, SO₃, Co₃O₄, CuO, MnO, and ZnO) are shown as a function of design test points using the KXT collector. The blue curves illustrate the variability in flotation recovery across the designed test series for each oxide.

Figure 13 presents the metallurgical recovery resulting from the interaction of operational variables in Test 3, which achieved an average recovery of over 46% of Co, Cu, and Zn oxides. However, the interaction in test 13 resulted in an average increase of over 53% for Co, Cu, and Zn oxides. On the other hand, MnO showed unsatisfactory recovery.

The metallurgical flotation tests (KXT1–KXT16) utilizing the KXT collector have demonstrated its efficacy in concentrating valuable minerals, notably copper (CuO) and cobalt (Co₃O₄), while concurrently rejecting gangue minerals such as silica (SiO₂) and iron (Fe₂O₃). The recovery of copper ranges from 29.55% to 82.57%, and the recovery of cobalt ranges from 56.08% to 82.57%, reflecting sensitivity to operating parameters such as pH, dosage, and grind size. Silica rejection is consistently strong, with 60%–80% of respondents reporting to tailings, while iron also predominantly reports to tailings (75%–95%), thereby minimizing contamination. Zinc (ZnO) is typically rejected to tailings, although exceptions (61.87% in KXT8) suggest occasional selectivity issues. However, magnesium (MgO) demonstrates notable recovery in both streams, indicating the buoyancy of its minerals alongside copper and cobalt, thereby diminishing concentrate purity. Overall, the KXT collector demonstrates strong potential for copper and cobalt recovery and excellent silica rejection. However, further optimization is required to control variability, limit zinc misreporting, and suppress magnesium flotation. KXT10 conditions offer the most promising balance in this regard.

A comparative assessment of flotation performance reveals that the KXT collector demonstrates substantial yet variable recoveries of copper (CuO: 29.55%–82.57%) and cobalt (Co₃O₄: 56.08%–82.57%. The rejection of silica (SiO₂: 60%–80% to tailings) and iron (Fe₂O₃: 75%–95% to tailings) was effective, yielding high-grade concentrates. This was despite challenges with magnesium (MgO) co-recovery and occasional zinc (ZnO) misreporting (The concentrate contains 61.87% ZnO (KXT data). In contrast, the (DTP) collector demonstrated consistent flotation efficiency, with copper recoveries frequently exceeding 50% and occasionally surpassing 60%, along with cobalt recoveries over 50%. This validates its strong selectivity towards these metals. However, it recovered 44%–61% of iron oxide into the concentrate, thereby compromising purity [13]. While KXT demonstrates notable proficiency in gangue rejection and concentrate quality under optimal conditions, DTP exhibits more consistent performance in Cu, Co, and Zn across a range of conditions, albeit with elevated Fe contamination levels. Collectively, these findings suggest that the most effective strategy to balance recovery efficiency and concentrate purity may be the optimization or blending of collectors.

4.3.3. KXT Metallurgical Mass Balance

Figure 14 presents the metallurgical balance data from tests 3 and 13, detailing the synergistic interactions among operational variables. In these tests, an enrichment of 1.4, 2.5, and 2.6 for Co, Cu, and Zn oxides was measured in test 3, while test 13 reproduced enrichments of 1.4, 2.6, and 2.7, respectively, for these same oxides. Finally, MnO did not improve metallurgical recovery in either test. The figure below illustrates the equation behavior of the metallurgical balance in the two tests mentioned here.

As shown in the bar chart, KXT 13 outperforms KXT 3 in metallurgical recovery across a range of metal oxides. The most significant discrepancy is evident in the case of Manganese Oxide (MnO), where KXT 13 achieves a noteworthy 93.3% recovery, exceeding the 6.5% recovery of KXT 3. Similarly, KXT 13 exhibited superior performance with Cobalt Oxide (CoO), achieving a recovery of 65.8% compared to 29.6% for KXT 3. While KXT 13 continues to demonstrate higher recovery rates for Copper Oxide (CuO) and Zinc Oxide (ZnO), the discrepancy is less pronounced. The data obtained from the experiment indicate that KXT 13 is a more effective flotation reagent for the separation and recovery of the specific metal oxides under investigation.

4.3.4. KXT Pyrite Recovery

Figure 15 provides a summary of pyrite (Pv) recovery in flotation tests (KXT1–KXT16), showing the distribution of pyrite across the concentrate and tailings streams, along with the corresponding recovery efficiencies.

MLA tests showed that the material from the Tharsis Mine consists of approximately 65% pyrite (FeS₂) in its mineral composition. Based on stoichiometric data calculated from the masses and grades obtained in the tests, it was possible to estimate the percentages of pyrite mass recovered, as well as its recovery and loss in the reject, as shown in the figure below:

The MLA Pyrite % remains constant at 65.87% across all tests, indicating a fixed feed mineralogy, while the distribution between concentrate and tailings varies significantly depending on flotation conditions. The differential Pyrite in concentrate ranges from 9.88% (KXT1) to 25.43% (KXT14), while the corresponding differential Pyrite in tailings shows an inverse trend, from 55.99% down to 40.44%. This indicates the equilibrium between recovery efficiency and pyrite rejection.

As demonstrated in Figure 14, the recovery values of concentrate Pyrite range from 14.99% (KXT1) to 36.52% (KXT6). Higher recoveries are generally associated with reduced pyrite losses to tailings. However, a trade-off is evident: tests with high recovery (KXT6 and KXT7, >34%) exhibit lower pyrite rejection, as indicated by lost pyrite-to-tailings values of ~63–65%, compared to lower-recovery cases such as KXT1 (85.01%).

The data demonstrate that flotation synergism results in variable pyrite recovery, exhibiting a discernible trade-off between recovery and rejection. Tests such as KXT6 and KXT7 achieved the highest recoveries, but at the expense of higher pyrite presence in the concentrate. Conversely, tests like KXT1 and KXT10 favored pyrite rejection but with limited recovery. This balance underscores the significance of process optimization in the context of selective management of pyrite flotation behavior.

4.3.5. KXT Flotation Kinetic (FK): Metallurgical Recovery

Figure 16 presents the results of the flotation kinetics (fk), showing that the metallurgical recovery reached satisfactory values between 15 and 18 min.

The data in Figure 15 above is best understood through the graph below, which compares metallurgical recovery to mass recovery over time, specifically in relation to floatability.

As illustrated in Figure 16, the flotation kinetics of the KXT collector are presented. It was observed that both metallurgical recovery (MR) and mass recovery (mR) increased with the duration of the flotation process; however, the most significant increases in recovery occurred during the initial 12 min. Thereafter, recoveries approached a state of equilibrium, plateauing. Co₃O₄, CuO, and ZnO exhibited the strongest response to KXT, achieving a recovery rate of more than 85% within 21 min, consistent with their well-documented affinity for xanthate collectors through stable metal-xanthate complexation [14,15,16,17,18,19,20]. The rapid recovery profiles of these materials, with recovery times ranging from 6 to 12 min, indicate favorable kinetics and efficient adsorption. In contrast, Fe₂O₃ and SO₃ exhibited moderate recoveries (50%–60%), while MnO demonstrated a negligible response (<40%), thereby emphasizing its poor flotation propensity under the examined conditions. It is important to note that the increase in MR was not directly proportional to mR, with the latter reaching only approximately 50% at 21 min. This divergence indicates a high degree of selectivity, with KXT preferring the flotation of target sulfide-associated minerals over indiscriminate bulk recovery. These findings reinforce KXT’s robust performance in Co, Cu, and Zn systems, as previously documented [21,22,23,24,25,26], while underscoring its deficiencies in MnO flotation. The findings indicate that the optimal flotation window is 12–15 min, achieving a balance between rapid kinetics and minimal incremental recovery beyond this period [27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46].

A comparative analysis of the flotation kinetics of DPT and KXT collectors is presented herein.

It was demonstrated that the collectors’ sample collection was effective at targeting Cu and Zn. DPT demonstrated a slight preference for longer flotation times to achieve maximum recovery. KXT exhibited faster kinetics and pronounced selectivity, particularly within the first 12–15 min. In contrast, both collectors were found to be less effective for MnO and moderately practical for Fe₂O₃, thereby highlighting the limitations in recovering oxide or less sulfide-associated minerals. These comparisons show that while DPT and KXT are both suitable for Cu-Zn enrichment, KXT may offer advantages in kinetics and selective flotation, enabling optimization of flotation time to balance recovery efficiency and operational throughput.

• An additional flotation bench test

Table 3. Parameters for three additional tests, designated A, B, and C, were conducted to evaluate the recovery of pyrite.

Parameters	A	B	C
Top Size (µm)	−10	−10	−200
% Solids	20%	20%	20%
Temperature	24.0 °C	24.0 °C	24.0 °C
pH	9.5	9.5	9.5
pH Conditioning Time	15 min	15 min	15 min
pH reagent	CaCo3	CaCo3	CaCo3
Depressor	0 g/t	0 g/t	0 g/t
Collector DanaFloat 527E	1300 g/t	0 g/t	1300 g/t
Collector Aerophine 3418A	700 g/t	700 g/t	700 g/t
Collector k-Xanthate	0 g/t	1300 g/t	0 g/t
Collector Conditioning Time	5 min	5 min	5 min
Frother	0 g/t	0 g/t	0 g/t
Frother Conditioning Time	0 min	0 min	0 min
Air Flow Rate	6 LPM	6 LPM	6 LPM
Mix Frequency	1000 RPM	1000 RPM	1000 RPM

Note: Note that the depressant column is there to make the experimental design matrix complete. Even though no depressant was used (0 g/t) in the current experiments, this parameter was kept in the table structure so that it would work with future tests that use selective depressants such as sodium metabisulfite or dextrin.

presents the parameters for three additional tests, designated A, B, and C, conducted to evaluate pyrite recovery. The tests mentioned above were based on conditions from other industrial tests.

The key parameters of the flotation tests were meticulously controlled to ensure comparable results across the three experiments. The top particle size was −10 μm for Tests A and B, while Test C used a significantly coarser size of −200 μm. The slurry solids content was maintained at 20% in all cases, with tests conducted at 24.0 °C and pH 9.5, and conditioned for 15 min using calcium carbonate (CaCO₃) as the reagent. It is noteworthy that no depressor or frother was applied during the tests. The principal distinction among the tests resided in the collectors employed: Test A employed DanaFloat 527E (1300 g/t) and Aerophine 3418A (700 g/t), while Test B relied exclusively on k-Xanthate (1300 g/t). Test C replicated the collector scheme used in Test A. Collector conditioning was standardized at 5 min across all experiments, while the airflow rate and mixing frequency were consistently set at 6 L/min and 1000 RPM, respectively.

Tests A and B were conducted with the same fine particle size (−10 μm) but used different collectors. This likely facilitated a direct comparison of the collector combination in Test A (DanaFloat 527E + Aerophine 3418A) with the single collector in Test B (k-Xanthate) under otherwise identical conditions.

Tests A and C used the same collector types and dosages, but differed in particle size. Test A utilized fine particles with a diameter of −10 μm, while Test C employed significantly coarser particles with a diameter of −200 μm. The objective of this comparison is to evaluate the influence of particle size on the recovery of pyrite when employing the same chemical reagents.

b.

Metallurgical Recovery

Table 4. Chemical composition of the concentrate and tailings streams for three different separation experiments (A, B, and C).

Test	Product	Fe2O3, %	SO3, %	Co3O4, %	CuO, %	MnO, %	ZnO, %
A	Conc.	43.60	58.67	87.54	56.95	88.39	38.45
	Tailings	56.40	41.33	12.46	43.05	11.61	61.55
B	Conc.	53.67	64.57	87.38	67.99	88.89	49.03
	Tailings	46.33	35.43	12.62	32.01	11.11	50.97
C	Conc.	38.28	51.45	83.04	56.78	84.76	34.27
	Tailings	61.72	48.55	16.96	43.22	15.24	65.73

presents the chemical composition of the concentrate and tailings streams from three separation experiments (A, B, and C), focusing on the percentages of several metal oxides.

The flotation tests demonstrated significant variations in metal recovery under differing conditions. Test B achieved the highest recovery for most elements (Fe₂O₃, SO₃, CuO, MnO, ZnO), indicating optimal parameters for maximizing overall metal yield. Conversely, Test C gave the lowest recoveries, particularly for Fe₂O₃ (38.28%), likely due to conditions designed to depress iron oxides. Test A was performed between B and C, with cobalt and manganese oxides exhibiting high floatability levels (greater than 83% recovery). In comparison, zinc recovery remained below 50%, indicating its limited response to the flotation reagents. The high recoveries of both High Co and Mn are indicative of effective concentration, although reporting Fe₂O₃ could dilute the target metal grade. The results demonstrate that the process can efficiently produce a Co-Mn-Cu concentrate from Zn-Fe gangue, with Test B exhibiting the optimal overall performance.

The results of the additional tests showed that the interaction effect of the collectors and industrial practice conditions applied revealed that:

a

Reducing the top size resulted in a significant increase in pyrite content.

b

Metallurgical recovery had a greater effect on MnO recovery compared to the use of KXT. These data should be further investigated, as the MnO levels at the feed are economically insignificant (0.006%) and do not justify using the configuration in test A as a reference.

c

The enrichment information projects incremental but not significant increases for the oxides of interest in polymetallic ores associated with pyrite.

d

In summary, a more in-depth study is strongly recommended to confirm significant percentage gains, along with an economic evaluation study to justify this methodology.

5. Conclusions

This investigative study examined the efficacy of two collectors, KXT and DTP, in selectively recovering pyrite from a sulfide ore deposit in southwestern Spain under both acidic and somewhat alkaline conditions. The findings indicate that KXT exhibited superior selectivity and recovery at pH 9, consistent with the favorable surface-charge conditions determined by zeta potential tests. The DTP performed adequately, particularly in regions exhibiting significant concentrations of CuO. The research indicates that the electrochemical environment exerts a significant influence on the efficiency of the collector and the effectiveness of the separation process. Subsequent studies that integrate pure mineral testing with advanced surface analysis will facilitate a deeper understanding of the mechanisms underlying adsorption, thereby advancing the development of reagents for complex sulfide systems. The primary conclusions that can be drawn from this analysis are as follows:

a

Ore characteristics: The sulfide ore deposit under scrutiny is constituted primarily of pyrite, accompanied by finely disseminated Co, Cu, Mn, and Zn oxides. These oxides are closely associated with pyrite grains, thus supporting the strategy of targeting pyrite as the primary carrier mineral for valuable elements.

b

Surface Chemistry: Zeta potential tests demonstrated that augmenting the pH above 8.5 generates a stable ionic environment, thereby enhancing the adhesion of collectors to pyrite surfaces. In such cases, xanthate has been shown to display enhanced electrostatic and chemical interaction with pyrite, thereby demonstrating a superior flotation response.

c

Collector performance: In the present study, the performance of two collectors was evaluated. The two collectors in question were DTP and xanthate. The study found that xanthate provided superior performance in selectivity and metallurgical recovery. The experimental conditions of Experiment 13 are recommended as a reference for pilot-scale optimization.

d

Due to variations in pyrite content, the flotation concentrate is expected to contain approximately 25% floated pyrite by mass. Flotation tests resulted in the following oxide mass recoveries (Table 4): Co₃O₄ ~88%, MnO ~88%, CuO ~57%, and ZnO ~39%. The studied metals occur associated with pyrite grains. Estimated recoveries in the concentrate are: Co, 60%–70%; Mn, around 90%; Cu, 35%–45%; and Zn, 30%–40%. Significant variability is likely in flotation results.

e

Kinetic behavior: The optimal flotation times were established as 12 min for DTP and 18 min for xanthate. Xanthate demonstrated an overall recovery efficiency of over 50% for total metallic oxides, with MnO exhibiting a recovery of 37%.

f

The following limitations and outlook are to be considered: The low recovery of manganese oxide (MnO) is attributed to two factors. Firstly, the feed grade of MnO is low. Secondly, there is a possibility that MnO is weakly associated with pyrite. It is therefore suggested that targeted mineralogical and reagent interaction studies be conducted in future work.

This study demonstrates that mechanical flotation, when optimized with xanthate as the collector, can effectively recover pyrite and associated oxide minerals from complex multiphase ores. The findings provide a compelling rationale for expanding the scope to encompass pilot testing and the formulation of reagent schemes aimed at optimizing selectivity and recovery.