Assessment of Cobalt Recovery from Copper Tailings by Leaching with a Choline Chloride–Citric Acid Deep Eutectic Solvent: Effects of Pretreatment and Oxidant Use

Abstract

The accelerating global demand for cobalt, driven primarily by lithium-ion batteries, has intensified the search for alternative sources of supply. Mine tailings represent a promising secondary resource, particularly in regions with extensive mining histories such as Chile. This study evaluates cobalt leaching from copper tailings using a deep eutectic solvent (DES), choline chloride–citric acid (ChCl–CA), with controlled addition of hydrogen peroxide. The tailings were subjected to pretreatments (froth flotation, chlorination, and thermal roasting) and then leached with choline chloride–citric acid-based DES or H₂SO₄. Temperature, leaching time, and solid–liquid ratio were evaluated. Results show that roasting significantly enhanced cobalt recovery when followed by citric acid or DES leaching, reaching up to 100% Co recovery. Under optimized conditions, DES-based leaching was effective and selective in a polymetallic matrix and achieved recoveries comparable to or better than acid leaching without generating toxic emissions. Although flotation and chlorination had limited effects on overall recovery, the results demonstrate the viability of integrated and cleaner technologies for valorizing tailings that contain critical metals such as cobalt.

1. Introduction

The global energy transition—from a coal-based energy mix to one increasingly supported by nonconventional renewable energy sources—has significantly intensified the demand for critical raw materials. Among these, cobalt (Co) is strategic due to its role in rechargeable batteries, advanced alloys, and catalytic applications [1]. By 2024, electric vehicle (EV) batteries accounted for approximately 43% of global cobalt consumption, a share that is expected to continue rising toward 2030 [2,3]. In this context, mine tailings derived from the beneficiation of copper sulfide ores have gained attention as a promising secondary resource. These finely ground residues often contain appreciable amounts of strategic elements such as cobalt and copper. Although traditionally discarded, tailings present a distinct advantage: they are already pre-mined and comminuted, thus substantially lowering the energy requirements for subsequent processing [4]. In Chile, one of the world’s largest copper producers, the mining industry is facing growing challenges related to environmental management, social responsibility, and energy efficiency. Tailings management is central because these materials constitute both a long-term environmental liability and a potential source of critical metals [1,2]. The valorization of these wastes aligns with both circular economy principles and the country’s strategic interest in adding value to its mining sector [3,5,6]. Cobalt recovery from silicate-dominated tailings has been explored in strong-acid and organic media, and more recently using hydrated deep eutectic solvents (DESs). However, direct comparisons across these chemistries at identical temperature and S/L on low-sulfide matrices remain scarce, obscuring whether performance differences arise from leachate chemistry or from the solid matrix. In this study, we deliver a head-to-head screening at 60 °C of sulfuric acid, citric acid, and a hydrated ChCl:CA DES, with controlled H₂O₂ dosing, on a single tailing material [7,8]. This design allows us to (i) disentangle oxidant demand via leachate, (ii) expose and rationalize a non-monotonic particle-size effect (a middlings window with clear implications for milling), and (iii) discuss DES viscosity and highlight the potential of DES-based leaching as a cleaner and selective alternative to conventional acid processes. Together, these elements constitute the novelty of this work and define a concise process map to balance yield and selectivity under mild conditions, contributing to the development of sustainable strategies for cobalt recovery from mining residues in the Chilean context.

2. Materials and Methods

2.1. Tailings Origin and Characterization

Tailings originated from a copper flotation circuit processing a sulfide ore in the Coquimbo Region, Central Chile. Bulk material was sampled at the thickener underflow; it was homogenized and riffle-split. Mineral phases were identified by X-ray diffraction (XRD). Particle-size distribution was measured, and, for experiments, the material was classified using stainless-steel laboratory sieves into +335 µm, −335 + 250 µm, −250 + 150 µm, −150 + 75 µm, −75 + 45 µm, −45 + 32 µm, and −32 µm. Tests carried out with unclassified tailings (full size range) are denoted as +335 to −32 µm.

The chemical characterization of the tailings was determined by an Inductively Coupled Plasma (ICP-MS) spectrophotometer (PerkinElmer SCIEX ELAN 9000, Waltham, MA, USA) equipped with a SimulScan detector system and a 2.5 MHz quadrupole. The most abundant elements found in the tailings are shown in

Table 1. Concentration of elements in the tailings.

Metals	Co (ppm)	Cu (ppm)	Al %	Ca %	Fe %	K %	Mg %	Na %	Total S %
Concentration *	33.3	3145.1	6.88	4.47	8.56	1.52	1.71	3.05	0.58

* Method detection limits for reported elements are given in Table S1.

, where Si is intentionally not reported due to non-quantitative dissolution of silicates under the adopted digestion conditions.

2.2. Reagents and DES Preparation

All the reagents used for leaching (choline chloride, citric acid, sulfuric acid, and hydrogen peroxide) were purchased from Merck (Darmstadt, Germany) and were of analytical grade and used without further treatment. Solutions were prepared using deionized water.

For flotation, a polyglycol–alkyl alcohol frother (MatFroth 355 used as received) and a thionocarbamate collector (Matcol D-101) both from; Mathiesen Company (Santiago, Chile); commercial grade, used as received) were employed. Slurry pH was adjusted with calcium hydroxide, Ca(OH)₂ (analytical grade, ≥95% purity, from Merck (Darmstadt, Germany)).

Sodium chloride (NaCl, ≥99% purity, anhydrous from Merck (Darmstadt, Germany)), used in the chlorination pretreatment, was crushed in a mortar to a fine, homogenized product size.

Two deep eutectic solvents (DESs) were prepared with different citric acid concentrations. These solutions were formed by adding 20% w/w of deionized water and 80% w/w choline chloride (CC) plus citric acid (CA) 1:1 and 1:2 molar (M) ratios and agitated in an ultrasonic bath at 80 °C until the solution was homogeneous. Water addition accelerates the synthesis; mixtures without water, but otherwise identical, that were agitated for >16 h at 80 °C did not form a stable eutectic.

2.3. Tailing Pre-Leaching Procedures

Pretreatments were applied independently to the same batch of as-received tailings. Specifically, chlorination and roasting were conducted directly on the tailings, whereas flotation was evaluated as an alternative route to produce a concentrate from the same tailings. These pretreatments were not applied sequentially to one another. Thereafter, each resulting material—raw tailings, flotation concentrate, chlorinated tailings, and roasted tailings—was leached separately with sulfuric acid under the conditions specified.

2.3.1. Froth Flotation

The tailings were froth-floated to further concentrate cobalt and copper. Given the frequent Co–pyrite association (as Fe↔Co substitution), operating conditions were selected within typical pyrite/copper–pyrite flotation ranges [4,9]: pH 10, which lies at the upper limit of the pyrite flotation window with thiol collectors and is commonly applied around 9.5–10 in Cu–FeS₂ systems. Air flow was fixed at 7 L·min⁻¹, a value used in laboratory Denver-type cells for sulfide flotation, and frother/collector dosages were 30 and 35 g·t⁻¹, respectively—within published laboratory/plant ranges for polyglycol–alcohol frothers and thionocarbamate collectors in Cu-bearing sulfides [10,11]. Agitation was 600 rpm during conditioning and 750 rpm during flotation, consistent with guidance to use lower impeller speeds. Conditioning and flotation times were 3 and 12 min, respectively, aligning with common laboratory practices (2–3 min reagent conditioning; 10–12 min batch flotation) [12].

2.3.2. Chlorination

An acid–chloride agglomeration and humid-curing pretreatment (80 °C, 6 h) targeted cobalt. In chloride media, Co(II) forms stable chloro-complexes (e.g., CoCl⁺, CoCl₂⁰, CoCl₃⁻, and CoCl₄²⁻), facilitating cobalt solubilization during the subsequent oxidative leach. This strategy follows agglomeration and curing practices widely reported for sulfide ores, adapted here to increase chloride activity in the solid bed while preserving moisture for low-temperature reactions [7,13].

Sodium chloride dosage was set at 70% of the stoichiometric requirement for in situ HCl generation according to the reaction H₂SO₄ + 2 NaCl → 2 HCl + Na₂SO₄ (via NaHSO₄), computed from measured H₂SO₄ consumption. Once the agglomerates had formed, they were transferred to a sealed glass reactor equipped with a gas inlet and outlet. Saturated steam maintained a humid atmosphere; the outlet stream was bubbled through a trap containing a sulfuric acid solution (pH 1.5) to suppress the release of chlorinated gases. The reactor was immersed in a thermostated water bath at 80 °C for 6 h [8,14].

In our material, cobalt is expected to occur either in discrete Cu–Co sulfides (e.g., carrollite) or isomorphically substituted in pyrite; both modes benefit from chloride–oxidant systems that either form soluble Cu/Co chlorides or prevent sulfur passivation, thereby enhancing Co release [15].

2.3.3. Roasting

Tailings were oxidatively roasted at 600 °C for 2 h under flowing air (100 mL min⁻¹). For cobalt, the literature and thermodynamic studies show that cobalt-bearing sulfides (e.g., cobaltite; Co hosted in pyrite) under structural/phase changes at ~550–650 °C form oxides that enhance subsequent dissolution of Co²⁺ in acid. Selective sulfation-roasting routes used on Cu–Co concentrates transform Cu/Co into water/acid-soluble sulfates, while iron is fixed as hematite—an established rationale for testing thermal pretreatments before leaching, even when the present work uses oxidative (air) rather than sulfating conditions [16].

2.4. Leaching Test

DES leaching. Hydrated ChCl–CA DES leaching was conducted in a jacketed glass reactor at 30 or 60 °C for 8 h. The reactor was charged with 2.5 g of tailings, 20 mL of DES solution and 10 mL of deionized water (solid–liquid ratio S/L ≈ 83 g L⁻¹). Hydrogen peroxide (30 wt%) was dosed at 5.0 mL h⁻¹ throughout the test. The slurry was continuously stirred to maintain all particles suspended.

Citric and sulfuric acid leaching. Two acidic media were evaluated: citric acid (CA) at 1 M and 5 M, and sulfuric acid at 2 M. CA leaching was performed in a jacketed reactor at 60 °C for 4 h, using 2.5 g of tailings and 40 mL of CA solution (S/L ≈ 62.5 g L⁻¹); H₂O₂ (30 wt%) was added at 2.5 mL h⁻¹. Sulfuric acid leaching was conducted at 60 °C for 4 h on seven particle-size fractions; reactors were loaded with 15 g of tailings and 90 mL of 2 M H₂SO₄ (S/L ≈ 167 g L⁻¹), and H₂O₂ (30 wt%) was dosed at 2.5 mL h⁻¹.

Leaching of chlorination-pretreated tailings. Tests were performed in an 800 mL reactor at 25 °C for 4 h, charging 100 g of tailings and 300 mL of an H₂SO₄-adjusted solution at pH 1.5 containing 70 g L⁻¹ Cl⁻ (denote the initial concentration calculated from NaCl addition; typical ranges of 20–50 g·L⁻¹ Cl⁻ (Section 2.3.2) are documented for hydrometallurgical chloride systems) (S/L ≈ 333 g L⁻¹). Hydrogen peroxide (30 wt%) was added at 2.5 mL h⁻¹.

Leaching of flotation products. Concentrate or tailings obtained from flotation were leached in a 100 mL reactor at 25 °C for 4 h, using 10 g of solids and 30 mL of 2 M H₂SO₄ (S/L ≈ 333 g L⁻¹); H₂O₂ (30 wt%) was dosed at 2.5 mL h⁻¹.

This work was designed as a screening study under constrained sample availability; not all leachate–matrix combinations were tested. In particular, applying the full leaching matrix to the flotation concentrate is a priority for future work to enable direct benchmarking against raw tailings.

Sampling and analysis: Liquid samples were collected at the end of each test (and, for selected experiments, at predetermined times), filtered through 0.22 µm syringe filters to remove suspended solids, and analyzed by AAS.

Micrographs were acquired using a Thermo Scientific Quattro S scanning electron microscope (Thermo Fisher Scientific, Waltham, MA, USA). Secondary-electron imaging was performed with an ETD detector under high vacuum. EDS microanalysis was carried out with an UltraDry detector (30 mm², 129 eV at Mn Kα; Thermo Scientific). The image shown was obtained at 10 kV, WD = 10.2 mm, magnification ≈ 1500× (field of view HFW ≈ 276 µm, VFW ≈ 184 µm), resolution of 1536 × 1024 px, dwell time of 5 µs/px, and single-frame integration.

3. Results

3.1. Tailing Characterization

The XRD pattern (Cu Kα, 20–90° 2θ, step 0.020°), Figure 1, is dominated by framework silicates consistent with zeolitic phases. The most intense reflection at 28.07° 2θ (d = 3.179 Å) matches zeolite Beta (SiO₂ 010), while additional reflections in the 21–32° 2θ window are accounted for by zeolite and ferrierite (Si₃₆O₇₂) (e.g., 21.00–21.02° [Qz + BEA], 23.19° [FER + BEA], 23.71° [FER], and 24.44° [FER]). Quartz (α-SiO₂) is evident at 26.77° 2θ (101) and contributes, together with BEA/FER, to medium-angle overlaps such as 39.58° and 42.57° 2θ. Metallic copper (fcc Cu) is identified by its characteristic triplet at 43.31° (111), 50.25° (200), and 74.23° (220). No cobalt-bearing crystalline phases were detected by XRD. Given the low bulk Co concentration and the XRD detection threshold (~1–3 wt%), cobalt is likely present as finely dispersed or poorly crystalline forms not resolvable by XRD. Low-angle BEA/FER features (<15–20° 2θ) are not captured by the present scan range. Overall, the pattern points to a zeolite-rich matrix with quartz, minor fcc-Cu, and traces of Co-bearing phase(s).

Sulfide minerals were not detected above the XRD detection limit (~1–3 wt%); pyrite, if present, is therefore minor. Sulfate phases (e.g., gypsum/jarosite) were qualitatively identified. Taken together with the bulk sulfur content (Table 1), these observations indicate a predominantly oxidized sulfur inventory in the sample, with sulfides expected to be minor.

SEM imaging reveals angular-to-sub-angular particles with abundant platy fragments and widespread fines coating coarser grains. The finer fractions expose more reactive surface, consistent with an SEM-based geometric estimate: Ā = 0.164 µm², ECD ≈ 0.453 µm, and S⁻ ≈ 0.656 µm² per particle (assuming sphericity). Using ρ = 2.7–3.0 g·cm⁻³, this corresponds to an indicative SSA of ~4.4–4.9 m²·g⁻¹, which aligns with the observed increase in Co recovery at smaller sizes (Figure 2).

Fields of view are independent areas selected after a reconnaissance scan. Images are provided for morphology only; spot-EDS at higher magnification did not detect Co above the method’s detection limit, consistent with bulk Co at tens of ppm.

3.2. Leaching Results

Initial tests with H₂SO₄ established reference behavior for future experiments in this work. The results of this leaching are shown in Figure 3.

Some leaching was carried out utilizing classified sizes to determine whether there is a fraction with a higher proportion of cobalt concentration within the tailings. No significant differences in concentration were found, but the effect of particle size can be seen, which is directly related to particle liberation.

Because the tailings have a very low concentration of cobalt, it was proposed to carry out different pretreatments to tailings, utilizing the full particle size range, to increase the concentration of cobalt phases or transform the possible insoluble phases for subsequent leaching. The results are shown in Figure 4.

Among the pretreatment routes, roasting delivered the highest Co extraction; however, when assessed on a multi-criteria basis—(i) Co recovery and selectivity (Co/Cu, Co/Fe); (ii) operational intensity, proxied by temperature/energy demand; (iii) reagent footprint and downstream neutralization needs; and (iv) EHS/process considerations (off-gas management and solids handling)—its high energy requirement, gas generation, and limited practical throughput rendered it not viable within our screening framework. The other pretreatments did not yield sufficient improvements, so subsequent work proceeded with leaching the as-received tailings without pretreatment.

The last leaching tests were carried out utilizing deep eutectic solvent, based on the previous works which indicate that cobalt (specifically the oxides) can be leached and reach good efficiencies.

In the first tries, the leaching was carried out with just tailings and DES, but the high viscosity of the solution plus the mineral caused problems with the agitation, since the mineral was not mixed homogeneously, and made sampling complex. Therefore, it was decided to add water to reduce the viscosity of the solution and allow proper leaching and sampling. The results of these experiments are shown in Figure 5.

4. Discussion

4.1. Leachate Chemistry and Oxidant Demand

Among the solutions evaluated, the leaching agent dictated both cobalt dissolution kinetics and selectivity. Sulfuric acid (H₂SO₄) promoted rapid, proton-driven attack, reaching high Co recoveries at 60 °C within 4–8 h (Figure 3). This performance, however, coincided with a measurable dissolution of matrix phases (notably Fe-bearing species), reflecting the non-selective nature of strong acid at moderate temperature [17,18]. By contrast, citric acid (CA) alone displayed slower kinetics at 60 °C, but it benefited markedly from H₂O₂ addition, which enabled higher Co extraction under the same thermal window (Figure 4). This behavior is consistent with an oxidative step preceding complexation: once Co(II) is made available from finely dispersed hosts, citrate stabilizes it in solution and curbs hydrolysis/re-adsorption [18,19,20,21].

Finally, the hydrated deep eutectic solvent (DES) ChCl:CA (1:1–1:2), aided by H₂O₂, delivered near-quantitative Co recovery at 60 °C and 8 h at S/L = 50 g L⁻¹ (Figure 5). We attribute this to a combination of (i) organic complexation and (ii) favorable activity in the eutectic medium, with water addition lowering viscosity and improving mass transfer. Overall, oxidant demand was highest in the CA/DES systems and comparatively modest for H₂SO₄, informing reagent-use trade-offs between kinetics, selectivity, and consumables [22,23].

Although the stripping of cobalt from the DES phase was not experimentally performed in this study, it is well established that cobalt can be selectively recovered from DES-based systems using an aqueous phase containing weak organic or acids (HCl, H₂SO₄, Cyanex 272, and others). During this process, the aqueous solution acts as the stripping agent, promoting the transfer of Co(II) ions from the DES into the aqueous phase. This step also enables the regeneration and reuse of the DES, minimizing solvent loss and closing the process loop [24,25,26].

4.2. Matrix Effects and Selectivity (Cu/Fe/S)

Matrix composition constrained selectivity outcomes. Bulk chemistry and XRD indicate a silicate-dominated gangue with minor sulfides and no resolvable cobalt-bearing crystalline phases. This suggests that Co resides as finely dispersed or poorly crystalline species, aligning with the lack of EDS confirmation at spot level despite tens of ppm bulk Co [13,16,27,28]. Under H₂SO₄, Co extraction rose in parallel with Fe dissolution, pointing to partial non-selective attack of Fe-oxides/silicates that host or shield Co. In citrate-based systems (CA and DES) at 60 °C with H₂O₂, Co gains were accompanied by comparatively smaller Cu dissolution (Figure 4), consistent with complexation-controlled solubilization that favors Co under the tested conditions. The limited sulfide inventory rationalizes the modest benefit of chlorination/roasting within our window: improvements in Co extraction are dominated by leachate chemistry rather than by conversion of refractory sulfide hosts [9,29,30,31,32].

4.3. Particle Size: Liberation vs. Diffusion

Particle size exerted a first-order control on extraction. Co recovery generally increased as size decreased, reflecting higher specific surface area and improved exposure of Co-bearing sites. A non-monotonic dip was observed between intermediate size fractions (previously S1 → S2; standardized as [size range]), followed by recovery at finer cuts. We interpret this as a middlings window where Co-bearing domains remain partially locked within silicate aggregates; further grinding enhances liberation and restores the upward trend [18,20,26]. At the finest cuts, any additional gains can be tempered by mass-transfer limitations and slurry rheology—particularly in viscous media—due to aggregation and boundary-layer effects. In hydrated DES, the balance between viscosity reduction by water and surface-area gains becomes critical [22,33,34,35].

Compared with sulfation roasting routes that require a high-temperature step before leaching, our mild-temperature (60 °C) screening delivers high Co yields while avoiding off-gas management and energy-intensive pretreatment (Figure 5). Relative to bioleaching mini-pilot demonstrations (≈30 °C, ≈10 days), our chemical systems achieve similar Co recoveries within hours, with tunable selectivity via citrate/DES chemistry and controlled oxidant dosing. These outcomes are consistent with hydrated ChCl:CA DES performance at low temperature reported and extend that rationale to silicate-dominated tailings under an identical T and S/L benchmark (

Table 2. Comparison of routes to recover Co.

Authors	Matrix	Route/Leaching Reagent	Key Conditions	Co Recovery	Pros and Cons
Peeters 2020 [26]	LCO (LIB cathode)	DES ChCl:CA (2:1) + 35 wt% H2O; use of Cu/Al as in situ reducing agents.	40 °C, S/L = 20 g L−1; 900 rpm; with Cu and Al present 1 h	≈98% Co; Li ≈ 93%	High extraction at low T; “green” solvent; integrated separation route Viscosity/water to be controlled; need for metal reducing agents (depending on feed)
Mäkinen 2020 [36]	Sulfide tailings (pyrite-rich)	Bioleaching (Fe/S oxidants) in stirred tank (continuous-batch mode)	30 °C, 100 g L−1 solids, 10 L bioreactor; adapted consortium ≈10 d	≈87% Co (Zn 100%, Ni 67%, Cu 43%)	“Soft” reagents, low chemical consumption; scalable to sulfide tailings Slow (days), sensitivity to pH/Fe3+/jarosites; biological control required
Özer 2019 [37]	Old tailings (Lefke, Cyprus)	Sulfate roasting + Na2SO4 25% → leaching	700 °C roasting; Na2SO4 promotes selective sulfation Hours (roasting) + leaching	90.1% Co, 71.2% Cu	Improved selectivity; high Co in leachate High energy and gases; prior thermal stage; SOx/Na control in circuit.

).

5. Conclusions

Sulfuric acid leaching delivers high cobalt recovery, especially when temperature and time conditions are optimized. However, the environmental drawbacks associated with the generation of toxic gases such H₂S limit its applicability in sustainable processes.

In contrast, DES-based leaching shows promising selectivity and efficiency without gas generation under the conditions tested.

Key operational parameters, such as temperature, leaching time, leaching agent concentration, and solid–liquid ratio, had a significant influence on metal recovery, highlighting the importance of process optimization. In this framework, the best condition parameters were 60 °C, 8 h, 1:2 (DES), and 50 g/L, respectively.

Although pretreatments such as flotation and chlorination did not enhance Co recovery in this material, the consistent solubility of cobalt phases suggests that direct leaching strategies remain viable. Overall, hydrated ChCl–CA DES presents a cleaner and potentially more selective alternative to conventional acid leaching, meriting further investigation and scale-up.