Recovery Process for Critical Metals: Selective Adsorption of Nickel(II) from Cobalt(II) at Acidic Condition and Elevated Temperature

Effective and sustainable separation processes for critical metals, especially for the physicochemically similar elements nickel and cobalt in battery recycling, are of great interest in the future. Selective adsorption represents a highly potential process for this purpose. In this publication, a silica adsorbent functionalized with an amino-polycarboxylate derivate (HSU331) was investigated regarding the selective adsorption of Ni(II) in the presence of Co(II) in acidic solution (pH range at equilibrium 1.8–2.3) at elevated temperature. Comparable maximum equilibrium loadings ( q max ) for Ni(II) and Co(II) of 0.59 μmol(Ni(II)) · μmol(Ligand)-1 (18.3 mg(Ni(II)) · g(Adsorbent)-1), and 0.52 μmol(Co(II)) · μmol(Ligand)-1 (16.0 mg(Co(II)) · g(Adsorbent)-1), respectively, were achieved at T = 50°C in single-component experiments. Under competitive conditions, the Ni(II) loading remained constant at 0.60 μmol(Ni(II)) · μmol(Ligand)-1 (18.4 mg(Ni(II)) · g(Adsorbent)-1), while the Co(II) loading drastically decreased to 0.09 μmol(Co(II)) · μmol(Ligand)-1 (2.7 mg(Co(II)) · g(Adsorbent)-1) in an equimolar dual-component system. Calculated stability constants of 3 · 103 and 0.7 · 103 L · mol-1, respectively, for the formed metal ion complexes of Ni(II) and Co(II) onto the adsorbent HSU331, clarify the clear selectivity of the adsorbent towards Ni(II) in the presence of Co(II) even at elevated temperature (T = 50°C).


Introduction
Rising global warming requires an energy transition away from fossil fuel-based power generation to a sustainable generation from wind and solar energy.For the success of energy transition, electricity storage systems for wind power and photovoltaic systems or electric vehicles are mandatory.As a result, state-of-the-art lithium-ion batteries (LIB), whose mixed oxide cathodes consist of lithium (Li) and certain transition metals such as cobalt (Co) and nickel (Ni), will be increasingly demanded in the future.In 2015, 49% of the globally produced refined Co was already used in the rechargeable battery market [1] and the European Union (EU) classified this metal already as critical raw material [2].According to estimates, the application of Ni in batteries for electric vehicles will grow by 39% annually until 2025.At similar growth rates, this corresponds to a Ni requirement in 2030 of more than 50% of the current global Ni production [3].
From 2030 onwards, the EU thus demands material recovery rates from batteries, i.a. for Ni and Co of 95% each [4].Klimenko et al. postulate that an improvement of current global Co recycling rates from 30 to 50% is imperative to address an otherwise inevitable Co shortage by the middle of the century [5].Consequently, accelerated development of effective and sustainable recycling approaches for end-of-life Li batteries becomes an essential technological task.
Today, LIB recycling e.g. based on the Batrec, and Duesenfeld process, respectively, which combine mechanical and thermal treatments with hydrometallurgical methods, mainly acidic leaching in the presence of reductants [6,7].After leaching, Ni and Co normally exist as Ni(II) and Co(II) in an acidic aqueous phase, from which they are subsequently recovered by solvent extraction, and precipitation, respectively.Depending on the leaching agent employed, the recovery rates for Co(II) and Ni(II) range from approximately 80 to 99% [6,8,9].Thus, considerable amounts of up to 20% of these valuable metals get lost via wastewater.In view of the required economic and ecological sustainability of industrial processes, further treatment of such sewage effluents are expedient.Another typical industrial process is electroplating [10][11][12], where Ni(II) and Co(II) are removed from acidic wastewater during plating or at the end of the process [10][11][12][13].
Particularly, adsorption represents an advantageous elimination method for heavy metal ions in aqueous phase because of low operational costs, efficient separation at low concentrations, and an adjustable selectivity towards a specific target component [14,18,22,23].Adsorption selectivity can be achieved by surface modification of an adsorbent matrix, like porous silica gels, with target-directed ligands [24,25].These matrix materials offer adjustable porosities according to the selected process conditions, defined structures, and immense possibilities of specific surface functionalization [25,26].
Generally, amino-polycarboxylic acids in a nonimmobilized form show high complex stabilities with Ni(II) and Co(II) [27] and are consequently suitable for functionalization of adsorbents for heavy metal recovery.For instance, Repo et al. [28] showed high adsorption capacities for Ni(II) and Co(II) using silica gels functionalized with aminopolycarboxylic acids.A selectivity towards Ni(II) was suggested when utilizing chitosan [29,30] as matrix matter.Several other matrices, e.g.acrylonitrile-divinylbenzene copolymer (AN-DVB) [31] or silica polyamine composites [32], were functionalized with amino-polycarboxylic acid derivatives which showed high adsorption capacities for Ni(II) and Co(II).
In this article, we demonstrate the outstanding performance of a mesoporous silica gel functionalized with an amino-polycarboxylate derivate for the selective adsorption of Ni(II) from Co(II) from an acidic model laboratory solution (pH = 1-4) at elevated temperature (T = 50 °C), which represents characteristic conditions for Ni(II) and Co(II) containing industrial process waters [16,33,34].These aspects complicate the search of suitable adsorbents for integrated or attached recycling steps in industrial processes because adsorption is usually an exothermic process, which implicates that higher temperatures support desorption [35].Moreover, low pH values favor desorption of cations, here Ni(II) and Co(II) [24].
Equilibrium batch experiments were performed in order to investigate its adsorption behavior in single-and dualcomponent systems and the influence of different experi-mental parameters (temperature, pH, and various molar ratios) on the selective Ni(II) adsorption.Furthermore, a desorption method at 20 °C is presented.
This study provides substantial information for the development of a selective adsorption-based separation process as a basis for a sustainable Ni(II) and Co(II) recovery in lithium-ion battery recycling and for a process-integrated Ni(II) and Co(II) recovery step in electroplating.
2.2.Elemental Analysis of the Functionalized Silica Gel.The amount of functionalized ligands onto the silica surface (surface coverage F) was obtained by elemental analysis using a Vario EL Cube Elemental Analyzer (Elementar Analysensysteme, Langenselbold, Germany) with helium as carrier gas and sulphanilic acid (p.a., Merck, Darmstadt, Germany) as calibration standard.Each sample consisted of 10 mg adsorbent and was analyzed in triplicate.
Initially, the pH value of each sample solution was adjusted to 3.5 with 65 wt% nitric acid.For equilibrium achievement, the centrifuge tubes rotated in an overhead rotator (Sunlab, Mannheim, Germany) with 40 rpm for 24 h at 20 °C.Subsequently, an analogous series of experiments at T = 50 °C proceeded in a temperature-controlled heating cabinet.After 24 h, aliquots (V = 0:15 mL) of each sample solution were diluted with 5 mL 0.65 wt% of nitric acid and analyzed with the ICP-MS.The equilibrium pH value (pH eq ) in all sample solutions was measured in order to determine the ratio of released protons (n(H + , released)) per adsorbed metal ion (n(M(II), adsorbed)) during adsorption.This ratio represents the number of ligand dentates that coordinate one metal ion, which is a measure of the resulting complex structure and was calculated as follows: where c 0 ðMðIIÞÞ constitutes the initial and c * ðMðIIÞÞ the equilibrium concentration in the sample solution and pH 0 gives the initial pH value.
Based on the measured equilibrium concentrations compared to the initial ones, specific dynamic and thermodynamic equilibrium parameters (equilibrium loading, stability constant, enthalpy, entropy, and free enthalpy of formation) were calculated to clarify the underlying adsorption mechanism and to verify the selectivity of HSU331 (expressed by the selectivity coefficient and reaction engineering selectivity).
According to the measured experimental concentrations in μmol • L -1 , each corresponding molar equilibrium loading q * in μmol(M(II)) • μmol(Ligand) -1 yields as: where V represents the sample volume in L, m ad the adsorbent mass in g, and F the functionalization degree in μmol(Ligand)•g(Adsorbent) -1 (determined according to Section 2.2).
Usually, the adsorption of bivalent metal ions on functionalized surfaces proceeds according to the following equilibrium reaction: in which the M(II) ions and the adsorbent ligand L form ion ligand complexes.Considering the equilibrium concentrations (μmol • L -1 ) of the formed complexes c * ðMðIIÞ/LÞ, non-adsorbed metal ions c * ðMðIIÞÞ, and free ligands c * ðLÞ, the stability constant (K ML ) of the metal ion ligand complex was determined according to The enthalpy (ΔH 0 ) and entropy (ΔS 0 ) were derived from the linear van't Hoff equation [37], which describes the equilibrium position of a chemical reaction as a function of temperature: By plotting ln K ML (ordinate) against 1/T (abscissa) [38], ΔH 0 and ΔS 0 resulted from the slope, and the intercept of the linear curve, respectively.
Subsequently, the free enthalpy of formation (ΔG 0 ) according to the Gibbs-Helmholtz Equation (Equation ( 6)) substantiates the endergonic or exergonic character of the adsorption [37]: Two characteristic parameters, (1) the selectivity coefficient (α Ni,Co ) and (2) the reaction engineering selectivity related to the formed Ni/HSU331 complex (S Ni/L,L ), were chosen for discussing the selectivity of the adsorbent HSU331 towards Ni(II) and calculated from the results of the dual-component experiments as ratio of the stability constants of the formed Ni(II) and Co(II) complexes (1): and as ratio of the amount of formed Ni/HSU331 complexes (n * ðNi/LÞ) at equilibrium and the difference between the initial amount of the adsorbent's ligands (n 0 ðLÞ) and the amount of free ligands (n * ðLÞ) at equilibrium (2), which means that the denominator describes the total occupied ligands [39]:   in Section 2.4.Supernatants were discarded after 24 h and replaced by 5 mL of nitric acid with different pH values (1.0, 0.5, 0.0, and -0.3, respectively) for disclosing the optimal desorption conditions.After equilibration (24 h) and rejection of the acidic supernatants, two washing steps (V = 10 mL) with ultrapure water took place in rotating sample containers for 15 min at 40 rpm, followed by a second adsorption step.Aliquots of every adsorption, desorption, and washing step were sampled, diluted, and analyzed for dissolved Ni(II), and Co(II), respectively, according to Section 2.3.
The desorption efficiency (η DE ) was calculated appropriate to the following equation: where n d represents the desorbed and n a the adsorbed amount of the respective transition metal ion.1).The characteristic isotherm parameters q max (maximum adsorption loading at equilibrium) and K L (Langmuir equilibrium constant) were determined according to the Langmuir Modell [40]:

Results and Discussion
and are summarized in Table 2.At T = 20 °C, adsorption of Ni(II) led to the highest q max of 0.72 μmol(Ni(II)) • μmol(Ligand) -1 , and whereas Co(II) was adsorbed by one-third less (q max = 0:49 μmolðCoðIIÞÞ • μmolðLigandÞ −1 ) and a fourfold lower In comparison to our results (see Table 2, column 5), Repo et al. [28] reported Ni(II) and Co(II) maximum loadings (mg(M(II)) • g(Adsorbent) -1 ) in the same order of magnitude.However, contrary to our results, the difference between the maximum loadings for Ni(II) and Co(II) was ten times smaller than that in our study, indicating a lower selectivity potential at competitive adsorption conditions.
Taking into account the industrial process conditions (e.g.elevated temperatures) for the Ni(II) and Co(II) separation, the adsorption experiments were also performed at T = 50 °C.The results showed that the q max of Co(II) remained nearly unchanged (0.52 μmol(Co(II)) • μmol (Ligand) -1 ), while the maximum adsorption of Ni(II) decreased by 20% (0.59 μmol(Ni(II)) • μmol(Ligand) -1 ).Nev-ertheless, the q max and K L values for Ni(II) were higher than those of Co(II), that is HSU331 separated still more Ni(II) than Co(II) at elevated temperature.Overall, these results indicate a higher selectivity of HSU331 towards Ni(II), which will be discussed in more detail for competitive adsorption in the dual-component system.
3.2.1.Adsorption Mechanism.By virtue of (1) the observed discoloration of HSU331 after contact with the metal ion solutions and (2) the significant pH shift in the sample solutions by approximately two units into the strongly acidic range at equilibrium, it can be assumed that the adsorption occurred as an adsorptive ligand complexation.
The complexation of Ni(II) and Co(II) resulted in a blue, and pink coloration of HSU331, respectively.Generally, blue octahedral or tetrahedral high-spin Ni(II) complexes and pink octahedral high-spin Co(II) complexes are known from literature [41,42] supporting our assumption of adsorptive ligand complexation.Additionally, similar pink colored Co(II)/DTPA octahedral complexes, immobilized onto a melamine-formaldehyde resin were reported in [43].
Due to the chelating character of the non-immobilized amino-polycarboxylic acids, e.g.EDTA and DTPA, complex formation of free hexa-or octadentate ligands with Ni(II) or Co(II) ions usually generates octahedral complexes with high stability constants between lg K ML = 16:4 and 20.2 [27,42,44].The considerably lower stability constants determined for the HSU331 complexes, summarized in Table 3 (column 6), are more comparable to those of only tridentate or tetradentate ligands, like N-methyliminodiacetic acid (MIDA) and nitrilotriacetic acid (NTA) with Ni(II) (lg K ML ðNi/MIDAÞ = 8:7 ; lg K ML ðNi/NTAÞ = 11:5), and Co(II) (lg K ML ðCo/MIDAÞ = 7:5 ; lg K ML ðCo/NTAÞ = 10:3), respectively [27].These results demonstrate a non-optimal exploitation of the immobilized HSU331 ligands, likely due to steric hindrance [28,45,46].In general, metal ion chelate complexes described in the literature reveal constantly higher stability constants for Ni(II) than for Co(II) and, according to the Irving-Williams series, are explicable by a higher ligand field stabilization for Ni(II) in coordination with oxygen or nitrogen-containing ligands [27,42].This tendency was also observable for the stability constants (lg K ML ) of the corresponding HSU331 complexes determined in this study (Table 3, column 6).Overall, the lg K ML values are at least two times smaller and thus less stable than for the nonimmobilized complexes discussed above.Immobilization of the ligand molecules of HSU331 lowers their degree of motion, so that possibly more unstable octahedral complexes are formed in which some coordination sites of the metal ions remain occupied by water molecules [42,[47][48][49].
In particular, the pH of the equilibrated system is a decisive factor influencing K ML , which becomes clear from the underlying complex formation reaction: where M 2+ represents a bivalent metal ion, L the ligand, and n the variable number of released protons.Due to Equation (11), the release of one proton per coordination site 4 Adsorption Science & Technology descends the pH into the acidic range over time, thereby shifting the equilibrium of complexation to the left and decreasing the stability of the complexes [50].
During the adsorption experiments, this pH shift was evident.When only adsorbent HSU331 is dispersed into an aqueous phase with an initial pH = 3:5, the pH value in equilibrium state shifts to 2.7, which demonstrates the expected net acidity of the HSU331 ligands.After contact of HSU331 with the metal samples, the pH of Ni(II) and Co(II) samples shifted from pH 3.5 to pH 1.8, and 2.2, respectively, which underline the assumption that the adsorption was caused by adsorptive ligand complexation.The ratio of protons released per adsorbed metal ion corre-lates directly with the number of coordinated ligand dentates, namely for Ni(II) 3 to 4 and for Co(II) 2 to 3 (Table 3, column 5).This result additionally confirms the discussed weaker complexation of Co(II) ions.

Influence of Temperature on Adsorption.
The fact that at elevated temperature (T = 50 °C) the HSU331 adsorption performance, especially for Co(II) adsorption, remains almost unchanged (Table 2, column 4) is discussed on the basis of the estimated thermodynamic state variables enthalpy ΔH 0 , entropy ΔS 0 , and free enthalpy ΔG 0 of formation (Table 4).The negative ΔH 0 (-13.5 kJ • mol -1 ) of the Ni(II) complexation reflects the exothermic character of this  Table 2: Characteristic Langmuir isotherm parameters (K L , q max ) depending on adsorptive and temperature.
Adsorptive .By increasing the temperature, according to Le Chatelier's law, the reverse reaction proceeds more intensively, which consequently leads to a lower loading of the adsorbent HSU331 (see Figure 1).Conversely, the Co(II) complexation onto the HSU331 surface is weakly endothermic due to the positive enthalpy of formation (2.8 kJ • mol -1 ).
Overall, the more negative free enthalpies (ΔG 0 ) of formation for Ni(II) adsorption at T = 20 °C and T = 50 °C compared to the corresponding free enthalpies for Co(II) adsorption (see Table 4) indicate an elevated spontaneous nature of the Ni(II) adsorption onto HSU331.For both temperatures, this correlates well to the demonstrably higher HSU331 loadings with Ni(II) compared to the loadings with Co(II) (refer to Table 2, column 4).
The positive entropies for Ni(II) and Co(II) adsorption onto HSU331, presumably caused by the release of metal ion-coordinated H 2 O molecules throughout the complexation [38], signified a generation of chelate complexes onto the HSU331 surface, in which a multidentate ligand occupies at least two coordination sites of a central metal ion [27,42].As summarized in Table 3, the release of more than two protons per adsorbed metal ion implies an occupation of the same number of metal coordination sites by the ligands' dentates and a simultaneous release of at least two water molecules per adsorbed metal ion causing the observed entropy increase in the system (Table 4, column 3).
A significantly higher ΔS 0 for the Co(II)/HSU331 complex formation compared to that for the Ni(II)/HSU331 complexing was also found by Smith and Martell [27], who investigated similar non-immobilized amino-carboxylate complexes of Ni(II) and Co(II).

Competitive Adsorption.
The resulting isotherms for the competitive adsorption onto HSU331 at T = 20 °C as well as T = 50 °C in the dual-component system Ni(II) : Co(II) are depicted in Figure 2. Isotherm data points are divided into two parts: the first data point at low initial concentration (c 0 ðMðIIÞÞ = 1,000 μmol • L −1 for Ni(II), and Co(II), respectively; zoomed area in Figure 2) and the remaining five data points (c 0 ðMðIIÞÞ ≥ 2,800 μmol • L −1 for Ni(II), and Co(II), respectively).At low initial concentration, the loadings of HSU331 are comparable (0.13/0.17 and 0.14/0.16μmol (M(II)) • μmol(Ligand) -1 for Co(II)/Ni(II) at 20 °C, and 50 °C, respectively).In this case, the surplus of ligands (metal ion : ligand ratio 0.4 : 1) avoids a competition for the ligands between both metal ions.
At higher initial concentrations, the adsorption loading of HSU331 with Ni(II) increases steadily until saturation, while the adsorption loading of Co(II) decreased by around 80% with rising initial concentrations compared to that in the single-component system.This can be attributed to the onset of competitive adsorption from a metal ion : ligand ratio ≥ 1.
Table 5 demonstrates the relations between pH eq , released protons per adsorbed metal ion, and the stability constant K ML at T = 20 °C.For low initial concentrations (c 0 ðMðIIÞÞ = 1,000 μmol • L −1 for Ni(II), and Co(II), respectively), the stability constants K ML (Table 5, line 1-2) are comparable with those of each single-component system at a similar metal ion : ligand ratio (refer to Table 3, line 2-3).The ratio of released protons per adsorbed metal ion of 3.0 represents the average value of the ratios calculated for the single-component systems Ni(II) and Co(II) at the same concentrations and T = 20 °C, which implies that also in the dual-component system the respective determined number of ligand dentates (Ni(II): 3.65; Co(II): 2.54; see Table 3), takes part in the coordination of the metal ions.
From an initial concentration of greater than or equal to 2,800 μmol(Ni(II)/Co(II)) • L -1 (metal ion : ligand ratio ≥ 1), the stability constants of the Ni(II)/HSU331 complexes always exceed those of the Co(II)/HSU331 complexes (Table 5, column 6), that is Ni(II)/HSU331 complexes are preferentially formed in the presence of ligand deficiency.Additionally, the equilibrium reached at pH 2.0 is more favorable for Ni(II) than for Co(II) complexation as demonstrated for the single-component system (Table 3, column 4).Overall, the resulting K ML values for Ni(II)/HSU331 complexes in the dual-component system are comparable values to those calculated for the single-component system.The release of about two protons (1.8-2.4,see Table 5) signifies that only two ligand dentates are involved in Ni(II) complex formation.This result theoretically allows the separation of a 1.5 to 2-fold higher number of Ni(II) ions compared to the single component system (released H + : adsorbed metal ion ratio 3-4).In practice, only an increase of 18% could be achieved, reflected by the slightly higher q max of 0.85 μmol(Ni(II)) • μmol(Ligand) -1 at T = 20 °C and an initial concentration of 15,000 μmol • L -1 (Figure 2), compared to 0.72 μmol(Ni(II)) • μmol(Ligand) -1 for the singlecomponent system (see Table 2).
In general, the temperature rise of 30 °C has no impact on the adsorption performance of HSU331 towards both metal ions (refer to Figure 2).However, the remarkable decrease of the maximum Ni(II) loading at T = 50 °C (Figure 2) and an initial concentration of 15,000 μmol • L -1 can possibly be attributed to the strong exothermic character of the Ni(II)/ HSU331 complexation (refer to Section 3.2.2).
Additionally, an isotherm for the molar ratio of Ni(II) : -Co(II) of 1 : 3 was generated in order to investigate the selectivity of HSU331 towards Ni(II) under non-equimolar conditions (Figure 3).Interestingly, even if Ni(II) represents the minor system component, still relative high maximum  Table 5: Correlation between proton release per adsorbed metal ion during adsorption, stability constant K ML , and molar equilibrium loading q * at T = 20 °C in different competitive adsorption experiments at Ni(II) : Co(II) molar ratio of 1 : 1.
In the experiments, where the total metal ion amounts were lower than the amount of ligands (Figure 3, first three data points), Co(II) was still adsorbed with a q max of 0.22 μmol(Co(II))•μmol(Ligand) -1 , and 0.44 μmol(Co(II))• μmol(Ligand) -1 at 20 °C, and 50 °C, respectively.The higher Co(II) loading of HSU331 at T = 50 °C can be explained by the endothermic character of Co(II) complexation (Table 4).However, with respect to a possible selectivity of HSU331 towards Co(II) at 50 °C, it should be noted that despite the discussed higher loading for Co(II) at this temperature, the removal efficiency for Ni(II) reaches 96%, and for Co(II) only 52%, respectively.
The decrease of the Co(II) loadings (Figure 3, data points with a metal ion : ligand ratio > 1) can be explained by the replacement of central Co(II) ions in the Co(II)/HSU331 complexes through Ni(II) ions, as already discussed for the equimolar system.
Calculated selectivity coefficients (α Ni,Co ) (refer to Equation ( 7)) strongly varied between 0.8 and 173 without any tendency regarding molar ratio, initial metal ion concentration, and temperature.All selectivity coefficients > 1 indicated a selectivity of HSU331 towards Ni(II).Comparable selectivity coefficients at T = 20 °C between 37 and 129 were also found by Repo et al. [29], who examined several amino-polycarboxylic acid ligands bound to chitosan.
Since the calculated selectivity coefficients showed no trend, the reaction engineering selectivity related to the formed Ni(II)/HSU331 complex (S Ni/L,L , refer to Equation ( 8)) was used instead as a parameter (Table 6) for evaluating the exploitation of the occupied ligands for the Ni(II) adsorption.
In the equimolar system, S Ni/L,L > 0:5 signifies a selectivity of the adsorbent HSU331 towards Ni(II), because more than half of the occupied ligands are coordinated with Ni(II), while at a molar ratio of 1 : 3, a S Ni/L,L value > 0:25 indicates the Ni(II) selectivity of adsorbent HSU331.

Desorption Studies.
In order to find suitable desorption conditions for Ni(II) and Co(II), nitric acid with varying concentrations of 0.1 to 2 M (pH 1 to -0.3) was investigated as desorption solution (DS).Desorption efficiencies for both metal ions depending on the acidity of the DS are presented in Figure 4.The application of a DS with a pH ≤ 0:5 ðcðHNO 3 Þ ≤ 0:32 MÞ led to a complete desorption of Ni(II) (η DE ~100%), whereas 0.1 M nitric acid ðpH = 1Þ was only able to desorb 50% of Ni(II).In contrast, Co(II) was almost completely released from HSU331 already at pH = 1.These results correlate with the discussed higher stability constants of Ni(II)/HSU331 complexes at different pH values (see Section 3.2.1).Moreover, it is most likely that the protonated HSU331 ligands represent the more stable molecular form, since, e.g.non-immobilized EDTA molecules predominantly exist as protonated H 5 EDTA + molecules in aqueous solution at pH = 0:5 [51].
Similar findings were reported by Repo et al. [28].They demonstrated that Ni(II) was still adsorbed at pH = 1:0 but Co(II) only marginally by amino-polycarboxylic functionalized surfaces.Additionally, a complete desorption of Ni(II) and Co(II) was accomplished with 2 M nitric acid.Our results prove that already 0.1, and 0.32 M nitric acid are sufficient for the selective desorption of Co(II), and Ni(II) from HSU331, respectively.As the adsorption performance of HSU331 for Ni(II) and Co(II) (data not shown) was maintained at a second adsorption after desorption, it can be assumed that a complete regeneration of the adsorbent was achieved.

Conclusion
Our experimental results evidence that the investigated adsorbent HSU331 selectively separates Ni(II) from Co(II) at 20 °C and 50 °C in acidic aqueous solution (pH < 3:5).An approximately seven-to nine-fold higher loading of HSU331 was obtained with Ni(II) compared to Co(II) due to the higher stability constants for the proposed Ni(II) chelate complexes.Low equilibrium pH values of 2.0 to 2.2 in the dual-component system reinforce the selectivity for Ni(II)/HSU331 complexes, while the Co(II)/HSU331 complexes exhibit significant lower stability constants (lg K ML ≤ 1:68).
Determined positive entropies of formation indicated that the generation of metal ion chelate complexes during adsorption is very likely.The thereby accompanied decrease of free enthalpies at T = 50 °C demonstrate that a temperature increase by 30 °C does not limit the metal ion adsorption.
Furthermore, based on our results from the desorption experiments, a two-step desorption process can be developed in which the selectivity of the adsorbent HSU331 is maintained.In terms of the recovery of both critical metals, the achieved results demonstrate the outstanding performance of HSU331 for a development of selective and thereby sustainable industrial processes for nickel and cobalt separation.

Figure 3 :
Figure 3: Adsorption isotherms of HSU331 for the dual-component systems with a molar ratio of Ni(II) : Co(II) 1 : 3 at T = 20 °C and 50 °C.Arrow marked data points denote the limit values from which the metal ion : ligand ratios are ≥1.Each isotherm consists of six data points.

Table 1 :
Initial concentrations of metal ions in the prepared singleand dual-component systems (molar ratio Ni(II) : Co(II) of 1 : 1 and 1 : 3) for the different adsorption experiments.

Table 3 :
Correlation between proton release per adsorbed metal ion during adsorption and stability constant K ML at T = 20 °C in a singlecomponent system.

Table 6 :
Selectivity of reaction engineering (S Ni/L,L ) for the Ni(II) adsorption onto HSU331 in dependence of molar ratio, initial concentration, and temperature.