Adsorption of Cobalt onto Zeolitic and Carbonaceous Materials: A Review

Abstract

At present, cobalt belongs to what are called critical raw materials due to its scarcity and its economic importance. Cobalt is a crucial element in the development of new technologies and applications for decarbonization, with around 40% of cobalt consumption being used for rechargeable battery materials. Additionally, cobalt-based catalysts are used in the production of hydrogen fuel cells, and this element is also employed in the production of superalloys for aerospace and power generation industries. For this reason, it is imperative to increase cobalt recycling by recovering from secondary sources, such as decommissioned lithium-ion batteries. Among the technologies for cobalt recovery, adsorption is a reliable alternative as it allows its recovery even at low concentrations in aqueous solutions and is relatively low in cost. Among the potential adsorbents for cobalt recovery, this paper reviews two of the most promising adsorbents for cobalt recovery from aqueous solutions: zeolitic and carbonaceous materials. Regarding zeolitic materials, the maximum adsorption capacities are reached by FAU-type zeolites. In the case of carbonaceous materials, the actual trend is to obtain activated carbons from a wide range of carbon sources, including waste, the adsorption capacities, on average, being larger than the ones reached with zeolitic materials. Additionally, activated carbons allow, in many cases, the selective separation of cobalt from other ions which are present at the same time in the aqueous solutions such as lithium.

1. Introduction

Nowadays, metallic elements are ubiquitous in our daily lives. They are present in a wide range of products and applications, from electronic devices to medical applications, energy, home appliances or urban furniture. Among metallic elements, those belonging to what are called “Critical Raw Materials” (CRM) are especially important, due to their high market value and large supply risk. Figure 1 shows the last CRM developed by the European Commission in 2023 [1]. Some of them, apart from being critical, are also regarded as strategic (in bold in

Table 1. 2023 Critical Raw Materials according to the European Commission (strategic raw materials in bold) [1].

Aluminium/bauxite	Coking coal	Lithium	Phosphorus
Antimony	Feldspar	Light rare earth elements	Scandium
Arsenic	Fluorspar	Magnesium	Silicon metal
Baryte	Gallium	Manganese	Strontium
Beryllium	Germanium	Natural graphite	Tantalum
Bismuth	Hafnium	Niobium	Titanium metal
Boron/borate	Helium	Platinum group metals	Tungsten
Cobalt	Heavy rare earth elements	Phosphate rock	Vanadium
		Copper	Nickel

) because they are crucial for the development of the strategic and technological planning of the countries. The degree of criticality lies in their economic importance for future applications and risk of supply. According to another European Commission report [2], the cobalt demand to develop Li-ion batteries is expected to boost by 4.1 times in 2030 and by 10 times in 2050, while the lithium demand for the same purpose is expected to increase by 15 times in 2030 and by 42 times in 2050.

Although, traditionally, these minerals have been extracted from ores by means of mining processes, their increasing demand makes it necessary to look for alternative ways of supply from secondary sources. In line with the Sustainable Development Goals (SDGs), these alternatives should fulfil two main objectives: waste management and selective recovery of metals. This will positively contribute to the SDG 12th (Responsible consumption and production) goal, contributing to an efficient use of natural resources and reducing waste, and SDG 13th (Climate action), e.g., cobalt is essential in the electrification of mobility. For example, in this sense, red mud [3,4,5,6] and phosphogypsum [7,8,9] have been employed as secondary sources of rare earths, while e-wastes have been employed as secondary sources of rare earths [10] or other critical metals such as gallium, indium [10,11,12], lithium [13,14], or cobalt [15,16], within the concepts of urban mining as the alternative of traditional mining and a circular economy [17], as shown in Figure 1.

Cobalt is a versatile metal with a wide range of applications in various fields. It is used in the production of magnetic alloys, hard metals, catalysts, pigments, and batteries. Cobalt is also a critical component in the production of lithium-ion batteries. The percentage of cobalt used in lithium batteries varies depending on the type of battery and the manufacturer. However, it is estimated that around 40% of cobalt consumption is used for rechargeable battery materials, such as lithium cobaltate for lithium-ion batteries and cobaltous oxide for nickel-metal hydride batteries. Velázquez-Martínez et al. indicate that the cathode of a lithium-ion battery (LiCoO₂) may represent up to 27% by weight [18].

Cobalt is also an essential element in the development of new technologies and applications for decarbonization. For instance, cobalt-based catalysts are used in the production of hydrogen fuel cells, which are an alternative to fossil fuels, or can be used in the production of superalloys that are used in aerospace and power generation industries.

Pure cobalt is not found in nature, but in association with nickel and copper ores, as well as in sedimentary rocks, soils, and minerals such as cobaltite and erythrite. However, as its average concentration is 25 ppm, it is clearly necessary to look for alternative resources [19]. Today, some cobalt is produced specifically from one of a number of metallic-lustered ores; however, cobalt is more usually produced as a by-product of copper and nickel mining.

Many techniques have been employed in the literature to selectively recover metals from these secondary sources. The most remarkable ones are solvent extraction, pyrometallurgical processes, membrane separation, ion exchange, chemical precipitation, or adsorption [20]. Among them, adsorption presents specific advantages such as the possibility of working with low metal concentrations (which is especially important when dealing with metal recovery from secondary sources), and the relatively low cost, depending on the prices of the adsorbent and its durability [21]. Adsorption allows continuous operation, with the possibility of recovering the adsorbent material by regeneration, and its reuse in successive cycles. Not only that, but adsorption has been used for water remediation and metal removal since those strategic metals could appear in leachates from incorrect disposal of the wastes in landfills.

2. Cobalt Adsorption

Adsorption, as the unit operation considered here, is the process in which a solute (adsorbate) dissolved or suspended in a fluid phase is retained in the surface of a solid material denoted adsorbent. According to the nature of the interactions between the adsorbate and the solid, the process can be classified as physisorption or chemisorption.

This technology has been deeply employed to recover a wide range of metals from aqueous streams. However, due to its special importance, this review will focus on cobalt.

Cobalt is a transition metal whose importance focuses on its technological applications, such as the production of superalloys with nickel and aluminum, powerful magnets, electroplating, pigments and dyes, and the manufacture of batteries. For instance, in the lithium-ion batteries, cobalt represents 5–20% of the composition, being a cathode metal, which increases the conductivity and energy density [22,23]. Cobalt has been selected as CRM in the 2023 evaluation, because of its critical applications previously described, plus its risk of supply, since the production of cobalt is focused on the Democratic Republic of Congo, and the mining operation in that country is associated with child exploitation and rights violations [24]. It is important to note the fact that, approximately, only 6% of the overall cobalt production is obtained as primary mine product [25], the rest being obtained as a by-product in the extraction of copper, nickel, and arsenic [26]. Additionally, according to the European Commission, the current recycling rate of cobalt is 22% [1], which is relatively low.

Because of the mining processes, the amount of cobalt in wastewaters from leaching processes can be greater than the recommended value of the environmental agencies. For example, although the concentration of cobalt in surface and groundwater in the United States is generally low, between 1 and 10 parts of cobalt in 1 billion parts of water (ppb) in populated areas, this value can be hundreds or thousands of times higher in cobalt-rich areas containing minerals or in areas near mining or smelting operations [27]. The maximum allowed concentration in cobalt in fresh water to protect aquatic life is 110 μg/L, with 4 μg/L of average concentration [28].

Consequently, the recovering of cobalt from these wastewaters as secondary sources is gaining importance. To fulfill this aim, among the previously described techniques, adsorption has been widely employed with a vast range of different solids. From this point on, this review will focus on two mayor groups of adsorbents that have centered on the removal of heavy metals such as this divalent cation, cobalt: zeolite and carbonaceous materials. The main advantages of zeolitic materials are their relatively low price, wide range of availability, and easy regeneration [29]. On the other hand, the main advantages of the carbonaceous materials are their relatively large surface area, that they can be prepared from a wide range of organic precursors (coal, wood, coconut shell, wastes, etc.), and that they can be prepared as mesoporous materials, which is what makes the adsorption process much faster [30].

3. Cobalt Adsorption onto Zeolitic Materials

3.1. Overview

Zeolites are crystalline and porous aluminosilicate minerals, with specific pore size. They can occur naturally in basaltic cavities, crystalizing due to hydrothermal alteration or diagenetic processes, but they also can be artificially synthetized. Their primary structure involves SiO₄ tetrahedra; if the Si⁴⁺ is the substituted by an Al³⁺, then an excess of negative charge is created, which must be compensated with intercalated cations; the different combinations of SiO₄ and AlO₄ tetrahedra lead to a wide range of three-dimensional structures [31]. Due to their structure, they possess quite interesting properties such as ion-exchange capacity, sorption capacity, shape selectivity, or catalytic activity. The sorption capacity is influenced by the size of the pore openings and the void volume, while ion-exchange selectivity is determined by the quantity and characteristics of cation sites and their accessibility. Additionally, catalytic properties are associated with pore openings, the dimensionality of the channel system, cation sites, and the space for reaction intermediates. [32]. The properties make zeolites suitable for retaining metal such as cobalt by means of ion-exchange processes. By way of example, the framework of the heulandite-type zeolite (HEU), a natural occurring zeolite, is shown in Figure 2 [31].

3.2. Cobalt Adsorption onto Zeolitic Materials

One of the first references is the work of Chmielewská-Horváthová and Lesný in 1992 [33], which employed natural mordenite and clinoptilolite as potential adsorbents. Although both zeolites showed relatively low capacities, mordenite resulted in being more suitable.

More recently, Erdem et al. [29] employed also clinoptilolite to remove cobalt with aqueous solutions with different initial cobalt concentrations (100–400 mg/L) and keeping the adsorbent dosage in 20 g/L at 25 °C. Under these conditions, the maximum cobalt removal attained was 77.96%, which was higher than the reached values for the other cations tested (66.10% for Cu²⁺, 45.96% for Zn²⁺, 19.84% for Mn²⁺). Moreover, Dávila-Rangel and Solache-Ríos [34] utilized two different clinoptilolites as well as kaolinite, as potential adsorbents of cobalt from aqueous solutions, and they studied the effect of the pH and contact time. According to their results, varying the initial pH between 4 and 8 did not affect the cobalt removal. At higher pH values, the authors argued that the adsorption capacity decreased probably due to cobalt hydroxide precipitation. Additionally, they found that both zeolites (commercially named Zefran and Zecrem) showed similar behaviors, with maximum adsorption capacities of 17.7 and 17.1 mg/g, respectively, while the maximum adsorption capacity of kaolinite was much lower. Qiu and Zheng [35] employed cancrinite-type zeolite with an Si/Al ratio near to 1, for lead, copper, zinc, and cobalt removal from aqueous solutions. The zeolite was prepared from fly ashes as precursors, using the molten-salt method, and its maximum adsorption capacities followed this order: Pb²⁺ > Cu²⁺ > Ni²⁺ > Co²⁺ > Zn²⁺. Mthombo et al. [36] synthetized ethylene vinyl acetate–clinoptilolite nanocomposites to remove copper, lead, and cobalt from aqueous solutions. They studied effects of the initial pH value and concentration of solutions, contact time, and solid dosage on the adsorption process, observing that, in all cases, equilibrium was reached after 24 h, the maximum adsorption capacities were reached at pH values between 5 and 7, and the adsorption selectivity trend was Pb > Cu > Co (in both single-metal and competitive adsorption experiments). The maximum equilibrium adsorption capacities were 0.9 mg/g for lead, 0.85 mg/g for copper, and 0.7 mg/g for cobalt, with an initial metal concentration of 20 mg/L in all cases. Seliman and Borai [37] investigated the affinity of natural chabazite and mordenite towards a cobalt, zinc, and nickel mixture, representative of industrial and radioactive wastes. For both zeolites, the affinity followed the order Zn²⁺ > Co²⁺ > Ni²⁺, but chabazite was more efficient than mordenite, achieving almost 100% removal of all the three metals in 120 min. The kinetics was so fast that within the first 10 min, already a removal of 74% cobalt, 91% zinc, and 54% nickel was achieved at 120 min. The authors attribute the initial fast adsorption to the ion exchange in the micropores on the surface of the zeolites’ grains, while the second slower stage to ion-exchange processes is attributed to the ion exchange in the micropores inside separate zeolites microcrystals. Other authors such as Gupta et al. [38] tested a hydroxyapatite/zeolite composite to remove cobalt from aqueous solutions. They studied the effect of pH, adsorbent dose, contact time, and initial cobalt concentration, and concluded that by employing the composite it was possible to increase the adsorption capacity of both zeolite and hydroxyapatite alone.

Zeolite AW-300 has also been employed [39] to remove cobalt from aqueous solutions. To analyze the efficiency of the process, the authors employed a 2³ experimental design, which determined that the process is influenced by the adsorbent dosage and initial cobalt concentration. With a zeolite dosage of 5 g/L, it was possible to reach a 100% cobalt removal efficiency from a 50 mg/L initial cobalt concentration. Another zeolitic material also used for cobalt removal was 13X [40]. This zeolite was employed to remove cobalt from monometallic aqueous solutions and from multi-metallic aqueous solutions containing cobalt, nickel, and chromium. The maximum monolayer adsorption capacities for Cr(VI), Ni(II), and Co(II) were 3.93, 6.19, and 10.39 mg/g, respectively, at 313 K. The authors also studied the re-usability of the solid by carrying out an adsorption–desorption cycle, and concluded that, by carrying out the desorption with 0.1 M HCl, it was possible to reuse the zeolite.

NaX, another zeolite with FAU structure, has also been used in the literature as a potential cobalt adsorbent. Deravanesiyan et al. [41] employed this zeolite to immobilize alumina particles and tested this material for cobalt and chromium removal from aqueous solutions. The authors showed that the immobilization of the alumina nanoparticles allowed the removal capacity to be increased by about 15% and 30% for Co (II) and Cr (III), respectively. Permutite, another synthetic zeolite, commercially available, has also been tested for cobalt removal [42], obtaining a maximum equilibrium capacity of 32.45 mg/g. Solache-Ríos et al. [43] employed a sodium-exchanged zeolite tuff mainly composed of clinoptilolite, mordenite, quartz, and anortite. According to the obtained results, the maximum cobalt adsorption capacity was 2.108 mg/g, at 303 K, with an adsorbent dosage of 10 g/L and an initial cobalt concentration of 20 mg/L. Ltaief et al. [44] synthetized an Na-faujasite type zeolite with an Si/Al ratio of 2.7, with illitic clay as precursor, reaching a maximum adsorption capacity of 93 mg/g, employing 0.5 g/L as the adsorbent dosage, and an initial cobalt concentration of 50 mg/L. The authors also carried out mono-metallic adsorption experiments with copper and chromium as adsorbates and checked that the selectivity order was Cu (II) > Co (II) >> Cr (III). The influence of the preparation method of zeolite was studied by Ansari et al. [45]. Employing an adsorbent dosage of 1 g/L and an initial metal concentration of 100 mg/L, it was possible to reach an adsorption capacity of 77.01 mg/g., at 25 °C, after 2 h. The authors claimed that the results achieved were comparable with other prepared zeolites but with the advantage of needing lesser preparation time. Mužek et al. [46] also employed X zeolite to remove cobalt from aqueous solutions, using a solid dosage of 5 g/L and four different initial cobalt concentrations (250, 350, 450, 590, and 700 mg/L). After 60 min at 298 K, the reached adsorption capacities ranged from 44 and 59 mg/g. More recently, Gulieva et al. [47] employed modified acid clinoptilolite by the saturation of the mineral grains with a 5% alcoholic solution of MEA to recover cobalt, copper, manganese, and zinc ions from aqueous solutions. The selectivity order that the authors found was Co²⁺ > Cu²⁺ > Zn²⁺ > Mn²⁺. A natural zeolite consisting of a mixture of clinoptilolite and mordenite was employed by Belova [48] to separate Cu, Ni, Co, and Fe ions from aqueous solutions. The author concluded that the sorption capacity of zeolite increased with metal ions’ concentration and that the selectivity order was Cu²⁺ > Fe²⁺ > Ni²⁺ > Co²⁺. Clinoptilolite was also employed by Rodríguez et al. [49] to remove cobalt ions from aqueous media. They compared the exchange capacity of bare clinoptilolite and compared it with the one of NaCl exchanged zeolite, concluding that the ion exchange process could increase the adsorption capacity by up to 21%. According to the authors, the optimum adsorption conditions to remove an initial cobalt concentration of 40 mg/L were 333 K, a pH of 5.5, 12 g/L of a clinoptilolite dosage, 0.9 mm of particle size, and clinoptilolite exchanged with NaCl. Under these conditions, the adsorption capacity was 3.35 mg/g. Araissi et al. [50] removed cadmium, cobalt, and nickel ions from aqueous solutions, employing NaY faujasite. The maximum adsorption capacities they obtained in monometallic experiments were 91, 50, and 54 mg/g for Cd²⁺, Ni²⁺, and Co²⁺ ions, respectively. Hossein et al. [51] increased the adsorption capacity of bare clinoptilolite by modifying it with magnetite particles and optimized the process with an experimental design. Under the most favorable conditions (pH 10, cobalt concentration 30 mg/L, adsorbent dosage 90 mg/L), the maximum removal capacity was 95%.

Cobalt and lithium are two cations that frequently appear together in the leaching of battery residues. For this reason, their sequential separation from aqueous media is interesting. Conte et al. [52] employed dolomite and mesoporous-activated carbon to selectively separate cobalt ions and, subsequently, 13X zeolite to recover lithium ions. Wang et al. [53] demonstrated that NaA zeolite showed high Co²⁺/Li⁺ separation selectivity, reaching > 95% Co²⁺ removal, while all Li⁺ remained in the solution. Additionally, Díez et al. [54] employed two FAU zeolites (13X and NaY) which were capable of achieving 100% cobalt removal, while lithium removal was about 30% (3.33 separation factor), with an adsorbent dosage of 5 g/L and initial Co and Li concentration of 40 and 20 mg/L, respectively. On the other hand, Hong et al. [55] employed NaX zeolite obtained from rice husk ash to adsorb cobalt. They observed, as previous authors, that the adsorption capacity was highly pH dependent, and they obtained the maximum adsorption capacity under these conditions: pH 3.0, 100 mg/L of Co(II) initial concentration, and 5 g/L of zeolite NaX. Finally, Nakhaei et al. [56] employed clinoptilolite to remove some heavy metals, cobalt among them, from synthetic aqueous solutions. They obtained that the adsorption capacity increased with the initial metal composition, and that the clinoptilolite was more efficient at adsorbing lead and cadmium than cobalt.

To end this chapter, a summary of the most representative studies analyzed in the review for the cobalt adsorption onto zeolites is presented in

Table 2. Comparison of cobalt removal adsorption capacities of different zeolites.

Adsorbent	Dosage (g/L)	Initial Concentration (mg/L)	qmax (mg/g)	pH	Equilibrium Time (min)	Reference
Zefran	20	100–400	0.60	4–7	1200	[34]
Zecrem	20	100–400	0.58	4–7	1200	[34]
Ethylene vinyl acetate-clinoptilolite	N/A	20	0.7	7	1440	[36]
Zeolite AW-300	10	50	10.39	5.7	1440	[39]
NaX	10	20	19.96	5–6.5	1440	[43]
NaY-faujasite	0.5	50	92	5.5	60	[44]
Clinoptilolite	12	40	2.9	5.5–6	45	[49]
13X	5	40	23	9–10	20	[54]
NaY	5	40	24.2	8–9.5	20	[54]

. As it can be observed, the maximum adsorption capacities are reached by FAU-type zeolites (92 mg/g) followed by a second group of zeolites (AW-300, 13X, NaX, NaY) achieving intermediate adsorption capacities (10–24 mg/g), and the least efficient group of zeolites (clinoptilolites and derived) (Zefran, Zecrem, Clinoplite and Ethylen vinyl acetate-Clinopolite) with low-adsorption capacities (0.5–2.9 mg/g). The large maximum adsorption capacity of the composite NaY–faujasite is noticeable. This can be justified by the lower initial cobalt concentration.

4. Cobalt Adsorption onto Carbonaceous Materials

4.1. Overview

Carbonaceous materials have been purposely employed as adsorbents due to their large surface area, their porous structure, and their superficial chemistry. This last characteristic is what determines their suitability as metal adsorbents: the presence of acidic groups on their surface allows the adsorption via ion exchange, although metals can also be adsorbed on the aromatic rings by electrostatic interactions with π electrons. Among them, activated carbons are the most employed materials, although other carbonaceous materials such as carbon nanotubes have also been used as adsorbents. In the literature, the employed activated carbons have been commercials, but also obtained from organic raw materials like wood, shells, stones, sawdust, peat, and different waste. In this case, it is necessary to apply a carbonization process to convert the cellulosic internal structure into a purely carbonaceous material [57]. Finally, the activated carbons are usually brought under an activation process to enhance the surface area and increase the number of oxygen-containing groups such as carboxyl, carbonyl, esters, lactones, and phenolic hydroxyl groups, and others, such as amido, sulfhydryl, acetamido, and amino [58]. The acidic groups introduced during the functionalization process have an affinity for complexation with metals and ion-exchange capacity [59]. The scheme of typical functional groups found in activated carbon is represented in Figure 3.

4.2. Cobalt Adsorption onto Commercially Activated Carbon

One of the first references of a carbonaceous material employed to remove cobalt ions forming aqueous solutions is the work of Netzer and Hughes in 1984 [60], who employed commercial granulated activated carbon for copper, lead, and cobalt adsorption, reaching removal percentages up to 90%. They performed selectivity studies and found that the activated carbon was more selective towards copper or lead than towards cobalt. Koshima and Onishi [61] employed commercially activated carbon from Merck to adsorb several metal ions (cobalt among them). Employing an adsorbent dosage of 1 g/L and initial cobalt concentrations between 0.1 and 10 mg/L, the percentage of adsorbed cobalt was more than 95%. However, similar percentages were reached for the rest of the ions tested (Cs, Y, Ce, Ti, Zr, W, Cr, Mn, Fe, Co, Ni, Ru, Cu, Ag, Zn, Cd, Al, Pb, Sb, and Bi) so the sorbent was not selective to cobalt. Paajanen et al. [62] employed different types of commercially activated carbon to remove cobalt from aqueous solutions. The adsorbents were prepared from different raw materials (coal, coconut shell, and peat). The adsorption tests were carried out with an adsorbent dosage of 10 g/L, with the particle diameter of the solid being between 0.14 and 0.30 mm, in buffered solutions. In the best conditions, the cobalt distribution coefficient between the liquid and the solid was over 100. More recently, Chen and Lin [63] employed an H-type commercially activated carbon, as a potential copper, zinc, and cobalt adsorbent. With a dosage of 1 g/L, this solid was capable of completely removing copper in about 5 h, but zinc and cobalt removal was slower. Khope et al. [64] employed a commercial granulated activated carbon containing adsorbed 3,5-dinitrosalicylic acid as the potential cobalt adsorbent. After equilibrium studies, they obtained adsorption capacities between 4 and 5 mg/g at 298 K. Recently, Hete et al. [65] tested two different grades of commercially activated carbon, to remove cobalt ions from aqueous acid solutions prepared from cobalt sulphate and with 3,5-dinitrobenzoic acid to acidify the medium. In light of the results, the authors concluded that cobalt is inaccessible to the inner pores of the carbons, so it is adsorbed superficially, and that the process of adsorption is surface-diffusion controlled. This previous study was further completed by employing other different commercial carbons [65] but reaching similar conclusions (granular-activated carbons are suitable adsorbents for cobalt removal). In a later work, Gunjate [66] employed an activated carbon chemically modified with 1,2-dyhydoxibenzene as complexing agent. The author concluded that the addition of this agent clearly improved the adsorption capacity of the solid. Kakavandi et al. [67] tested granulated activated carbon modified with sodium dodecyl sulfate as cobalt adsorbent and analyzed the influence of several variables such solution pH, contact time, adsorbent dosages, initial metal ion concentrations, temperature, and agitation speed on the adsorption efficiency. They concluded that, under optimum conditions, it was possible to completely remove an initial cobalt concentration of 20 mg/L, with an adsorbent dosage of 1.2 g/L. Finally, Gunjate et al. [68] adsorbed cobalt employing a commercial granular activated carbon modified with 3-aminophenol and 2-hydroxy-5-methoxy benzoic acid and concluded that the material chemically modified with 2-hydroxy-5-methoxy benzoic acid showed better efficiency towards cobalt adsorption.

4.3. Adsorption of Cobalt onto Activated Carbon Derived from Various Organic Wastes

In the literature, many alimentary residues have been employed as adsorbents for the recovery of metals. However, in this case, there is a problem to be considered: the regeneration of the adsorbent must usually be carried out in acidic conditions. To overcome this drawback, one possible solution is employing these residues not directly as adsorbents, but as active carbon sources.

One of the first references which synthetized activated carbon from organic raw materials to adsorb cobalt used almond shells as a carbon source [69]. The authors activated the carbons physically, employing carbon dioxide and, additionally, and also chemically with some of the samples with nitric acid. They concluded that increasing the activation time or adding a chemical activation step resulted in larger adsorption capacities. In this last case, it was possible to reach a 100% adsorption removal. Later, Srivastava et al. [70] employed waste slurry from fertilizer plants to produce a carbonaceous material by means of oxidizing treatment with hydrogen peroxide followed by a physical activation with air at 450 °C. The authors observed that cobalt (II) was poorly adsorbed in comparison to other ions such as chromium or lead. Teker et al. [71] employed activated carbon from rice hulls as a potential cobalt adsorbent. Different variables such as pH, adsorbent dosage, contact time, initial adsorbate concentration, and temperature were investigated, with the optimum values being pH 6.7, 30 g/L of solid dosage, and 40 min as adsorption time. By employing an initial cobalt concentration of 15 mg/L, the authors obtained cobalt removal efficiencies between 80% and 90%. More recently, Demirbas [72] employed hazelnut shells to prepare an activated carbon as a cobalt adsorbent. This author studied the influence of several variables, such as initial cobalt concentration (13.30–45.55 mg/L), mixing speed (50–200 rpm), pH (2–8), temperature (293–323 K), and carbon particle size (0.80–1.60 mm). In the best conditions, an adsorption capacity close to 4 mg/g for an initial cobalt composition of 13.30 mg/L, 298 K, pH 8, 150 rpm stirring rate, and 1.00–1.20 mm particle size was obtained. More recently, Krishnan and Anirudhan [73] employed bagasse pit as a source of activated carbon, for cobalt removal from aqueous solutions. To enhance the removal capacity, the carbon was sulfurized by means of a steam activation process, at 400 °C in H₂S and SO₂ atmosphere. The batch experiments they performed showed that the maximum removal capacity was achieved in the pH range of 4.5–8.5. Additionally, when using initial cobalt concentrations of 50 and 100 mg/L, the removal efficiencies were 90% and 81%, respectively (adsorption capacities of 45 and 20 mg/g). Arifi and Hanafi [74] employed activated carbons prepared from almond shells and activated on a carbon dioxide stream, but with different pore size distribution, to remove cesium, thallium, strontium, and cobalt radioisotopes from aqueous solutions. Employing a material whose mesoporosity is more developed (initial ion concentration = 0.006 mg/L, adsorbent dosage = 25 g/L), the authors reached a cobalt adsorption percentage of 50% (q = 1.18 × 10⁻⁴ mg/g). Salas-Tort et al. [75] employed coconut shells as a carbon source to obtain activated carbon capable of removing cobalt and nickel ions from aqueous solutions. They studied the influence of pH as well as of the initial metal concentrations, in batch mode experiments, reaching in all cases, recovery degrees up to 95% at room temperature (adsorption capacity of 4.7 mg/g for an initial cobalt concentration of 100 mg/L and an adsorbent dosage of 20 g/L). Prabakaran and Arivoli [76] prepared activated carbon from Thespesia Populnea in sulfuric acid medium. In batch mode, they studied several variables such as the initial cobalt concentration, pH, and adsorbent dosage. The authors concluded that pseudo-first order model is adequate to predict the kinetic of the process. Additionally, the authors carried out desorption studies in acid media, and concluded that hydrochloric acid is the best option for desorption as, with this acid, more than 90% removal of adsorbed metal ion takes place. Aljundi and Al-Dawery [77] used date seeds as the carbon source to obtain an activated carbon by pyrolysis in nitrogen atmosphere and further activation with carbon dioxide at three different temperatures (700 °C, 800 °C, and 900 °C). The authors observed that activating at 900 °C leads to an increase in the mesoporosity of the material. They carried out isotherm experiments where the isotherm data at different temperatures followed a type III adsorption isotherm model, which can be explained by Freundlich and BET models. The use of apricot stone as activated carbon precursor was deeply studied by Abbas et al. [78]. With this raw material, they prepared a chemically activated carbon with phosphoric acid. They analyzed the influence of several variables such as particle size, pH, initial cobalt concentration, and stirring speed, in bath mode experiments. They carried out both kinetic and isotherm experiments and concluded that the best models to describe both processes were pseudo-second and Freundlich models, respectively, with the maximum adsorption capacity calculated from the Langmuir model 111.11 mg/g at pH 9. More recently, Kyzas et al. [79] synthetized H₃PO₄ chemically activated carbon from waste potato peels and employed it as cobalt adsorbent from synthetic aqueous solutions. The authors studied the influence of several variables, such as pH or temperature. They found that cobalt removal was very small at acidic pH values, with being 6 the optimum pH value, in agreement with previous studies (with an initial cobalt concentration of 200 mg/L and adsorbent dosage of 1 g/L, they reached more than 87.5% cobalt removal, with an adsorption capacity of 175 mg/g). The increase in temperature led to an increase of the adsorption capacity.

Considering other different carbon sources, Zhang et al. [80] employed Xanthoceras Sorbifolia Bunge hull as a source of an activated carbon further employed to remove iron, cobalt, and nickel ions from aqueous solutions; they obtained maximum monolayer adsorption capacities of 241 mg/g for Fe(III), 126 mg/g for Co(II), and 187 mg/g for Ni(II), respectively. Additionally, they carried out regeneration studies and concluded that the activated carbon could be recycled up to three times for Co(II) and up to four times for Fe(III) and Ni(II) ions with very little loss of efficiency. Liu et al. [81] prepared an activated carbon from coconut husk as raw material and HCl and hydrogen peroxide as activated agent and used it to adsorb cobalt from aqueous solutions. They observed that increasing the activation temperature (from 25 °C to 75 °C) as well as the activation agent (from 0.01 to 0.04 M) led to an increase in the adsorption capacity. They also observed that H₂O₂ activated carbon was more capable of adsorbing cobalt (86% efficiency) than HCl activated carbon (79% efficiency). Kamble et al. [82] prepared activated carbon employing Ficus benghalensis (FB), Mangifera indica (MI), Tamarindus indica (TI), Azadirachta indica (AI), and Syzygium cumini (SC) plants as the carbon source. They observed that the maximum adsorption capacities were reached with 5 g/L (AI, MI) and 6 g/L (TI, FB, SC) of adsorbent dosage and 25 mg/L of cobalt initial concentration. Kolvankar [83] employed coconut tree root to prepare a carbonaceous material, and activated it with phosphoric acid, to obtain an activated carbon suitable for cobalt adsorption. The author studied the process by using a central composite design methodology and concluded that the three parameters with greater influence of the adsorption process were initial pH, solid dose, and initial cobalt concentration. In the best conditions (pH 6, cobalt initial concentration of 50 mg/L and 5 g/L of adsorbent dosage), the author reached a 100% cobalt removal and 10 mg/g adsorption capacity. Later, Ketsela et al. [84] employed White Lupine husk to prepare activated carbon and further employed it to adsorb lead, cobalt, and iron from aqueous solutions. The maximum removal efficiencies of Pb (II), Co (II), and Fe (II) obtained were 91.9%, 90.4%, and 90.3%, respectively. Citrus limetta leaves were employed to prepare activated carbon by Aboli et al. [85]. They employed this solid to adsorb lead, cobalt, and nickel from aqueous solutions, and the maximum adsorption efficiencies were reached at pH 6, with an adsorbent dose of 1 g/L and 5 mg/L of initial metal concentration. The obtained efficiencies were 99.5%, 98.6%, and 97.5% for Pb, Co, and Ni ions, respectively.

Continuing with different carbon sources, Lawan et al. [86] employed Detarium Microcarpum seeds to prepare an activated carbon to further remove cobalt and lead from aqueous solutions. They observed that the adsorption efficiency increased at higher adsorbent doses and metal initial concentrations. They also observed similar adsorption capacities of both metals, and that the adsorption removal efficiency increased with particle size (1400 µm > 420 µm > 150 µm). On the other hand, Gómez et al. [87] and Bernabé et al. [88] synthetized a mesoporous activated carbon using sucrose as carbon source, silica gel as the template, and diluted oxygen as the activating agent. They employed this material to remove cobalt from aqueous solutions and it was possible to remove 90% of the initial cobalt concentration under optimal conditions (298 K, 12.5 g/L, pH = 6). Additionally, it was possible to completely regenerate the solid with nitric acid and, by reducing the volume of the regenerative solution, it was possible to pre-concentrate the cobalt up to eight times. Later, this same material was employed by Conte et al. [52] to separate cobalt from lithium in a sequential process, and its synthesis was further optimized using an experimental design [89]. Such studies were continued by Conte and Gómez [90] by improving the activation process, through a combination of physical/thermal and chemical treatments with NaClO₂/H₂O₂ to increase the surface acidity. The double-activated mesoporous carbon obtained proved its efficiency in adsorbing high initial concentrations of Co²⁺, Ni²⁺ and Mn²⁺ (200 mg/L) using a dosage of 7.5 g/L and obtaining removal efficiencies over 80%, while at the same time, being able to separate selectively divalent and monovalent ions, since Li⁺ was barely adsorbed [65].

4.4. Cobalt Adsorption onto Other Carbonaceous Materials

Not only activated carbons have been employed as potential cobalt adsorbents derived from carbonaceous materials. In this sense, Ehsaninamin [91] employed oxidized multi-walled carbon nanotubes to remove copper, cobalt, and manganese ions from aqueous solutions. The author analyzed several variables, such as pH, solid dosage, and initial metal ion concentration, and found that the affinity order of the carbon nanotubes towards the ions followed the trend of Cu²⁺ > Co²⁺ > Mn²⁺. The experimental adsorption capacities were around 40 mg/g, 15 mg/g, and 7.5 mg/g for copper, cobalt, and manganese, respectively. More recently, Liu et al. [92] compared commercially activated carbon vs. commercial graphene as potential adsorbents of cobalt complexes from desulfurization and denitration wastewaters. After having studied several variables, such as pH or temperature, the authors found that the highest adsorption efficiency of graphene could reach more than 95% and activated carbon could reach more than 78%. More information about the heavy metal adsorption onto new carbonaceous materials can be found in Chai et al. [93].

Table 3. Comparison of cobalt removal adsorption capacities of carbonaceous materials.

Adsorbent	Dosage (g/L)	Initial Concentration (mg/L)	qmax (mg/g) (mg/mmol) *	pH	Equilibrium Time (min)	Reference
Commercially activated carbon	10	10	N/A	4	120	[60]
Commercially activated carbon	1	0.1–10	N/A	11.8	1200	[61]
Rice hulls	30	15			40	[71]
Hazelnut shells		13.30	13.88	6		[72]
F-200	2.5	0.5	2.75 *	5	360	[64]
F-400	2.5	0.5	2.13 *	5	360	[64]
Bagasse pit	2–15	50–100	22.58–40.50	6	240	[73]
F 200 D	2.5	N/A	0.0612 *	5	360	[65]
F 300 D	2.5	N/A	0.1059 *	5	360	[65]
Coconut shells	20	0.1–1	N/A	3	360	[75]
Apricot stone	5	80	111.11	9	20	[78]
Waste potato peels (PoP400)	1	10–1000	373	6	180	[79]
Waste potato peels (PoP600)	1	10–1000	405	6	180	[79]
Xanthoceras sorbifolia Bunge Hull	N/A	550–800	126.05	2.0–6.5	20–120	[80]
Sodium docecyl sulfate granular activated carbon	1.2	20	51	N/A	N/A	[67]
Mesoporous activated carbon	12.5	20	1.6	6	15	[88]
Mesoporous activated carbon	10	20	5.8	3.5	15	[89]
Mesoporous carbon doubly activated	7.5	200	22.3	4.9	15	[90]

The * indicates that the capacity is expressed in mmol. N/A means Non Applicable.

provides a summary of the most representative studies previously reviewed. As can be observed, the reached adsorption capacities are on average higher than in the case of zeolitic materials which, from our point of view, make these materials more suitable for cobalt adsorption.

5. Conclusions and Final Remarks

Cobalt belongs to a critical raw materials group due to its wide range of applications and its difficulty in extraction. For this reason, it is fundamental to look inside secondary sources for its recovery, as well as to analyze the different options to face this aim.

This paper focuses on the adsorption process for the recovery of cobalt from different sources. This technique has the advantage of the possibility of employing low initial metal compositions, as well as is relatively economical depending on the nature of the adsorbent. The first advantage is particularly important as the amounts of cobalt in leaching streams from secondary sources are in the range of ppm. Regarding the second advantage, this paper focuses on both zeolitic and carbonaceous materials, as these two groups include most of the employed solids as cobalt adsorbents.

Comparing both materials, zeolites are crystalline aluminosilicates with a definite porous structure. This allows for the selective adsorption of different metallic ions based on their size and shape. On the other hand, activated carbons are amorphous carbon materials with a highly porous structure that provides a large surface area for adsorption. In relation to adsorption capacity, while zeolites typically have a higher adsorption capacity for specific molecules due to their ordered pore structure and ion-exchange properties, activated carbons typically have a high surface area but may not be as selective as zeolites. Finally, when it comes to the desorption of adsorbed metals, the reported results are, in general, optimistic. However, as the desorption is usually carried out under acidic conditions, this may limit the usefulness of zeolites, due to their structure that could be altered. In the case of zeolitic materials, the maximum adsorption capacities are reached by FAU-type zeolites. On the other hand, in the case of the carbonaceous materials, it can be concluded that, although there are some references which employ commercially activated carbon as cobalt adsorbent, the trend is to synthetize activated carbon from different carbon sources, with the adsorption capacities, on average, being larger than the ones reached with zeolitic materials. Additionally, in many cases, it is possible to selectively separate cobalt from other ions, mainly lithium. Concerning the influence of several variables such as temperature and pH, the revised literature indicates that the influence of temperature on cobalt adsorption is not noticeable, and that the influence of pH is not significative at pH values between 3 and 8. Higher pH values could give rise to cobalt hydroxide precipitation.

In summary, after this study, it can be concluded that the main advantages of carbonaceous materials are their large adsorption capacity, that they can be obtained from a wide range of raw materials, that they can be obtained in a granular size (making easier their employment in continuous experiments), and that they have larger stability (which is an important advantage when adsorbing metals as the optimal working pH is in many cases acidic, and the desorption conditions also imply acid pH values). On the contrary, the zeolitic materials, on average, present better selectivity mainly due to their well-defined pore structure, although their use implies secondary contamination of water with aluminum ions or dissolution problems.

Finally, it is important to note that although nowadays this study has demonstrated that adsorption is a reliable alternative for cobalt recovery, future work must be focused on increasing the selectivity towards cobalt by either changing the working variables or modifying the adsorbents, as more and more complex matrices are generated.