Mechanism of Selective Extraction and Separation of Vanadium and Aluminum from Oxalic Acid Leachate of Shale: Experimental Investigation and DFT Calculations

Abstract

Oxalic acid serves as an environmentally benign leaching agent, exhibiting strong reducing and complexing capabilities. In the oxalic acid leachate derived from vanadium-bearing shale, aluminum ions are present as major impurities. Achieving efficient and deep separation of vanadium from aluminum remains a key technical challenge. This study investigates the selective separation of vanadium and aluminum from oxalic acid leaching solutions using solvent extraction with Aliquat 336, supported by density functional theory (DFT) calculations. Experimental results demonstrate that, under optimized conditions, Aliquat 336 enables effective separation of vanadium from aluminum. DFT analysis elucidates the molecular-level interaction mechanism, revealing that the binding affinity of Aliquat 336 for [VO(C₂O₄)₂]²⁻ (ΔG = −287.96 kJ/mol) is significantly stronger than for [Al(C₂O₄)₂]⁻ (ΔG = −186.68 kJ/mol). These results provide a solid thermodynamic basis for the observed selectivity and establish a robust theoretical framework for developing high-efficiency separation processes. This work thus clarifies, for the first time, the mechanistic foundation of vanadium–aluminum separation in oxalic acid systems.

1. Introduction

Vanadium is a strategically important metal, widely utilized in the production of alloy steels and catalytic materials, and is often referred to as the “monosodium glutamate of industry” due to its performance-enhancing role [1,2]. Vanadium-bearing shale represents a significant alternative resource for vanadium supply, complementing the more conventional source of vanadium titano-magnetite [3]. Driven by sustained demand from the steel industry and the rapid expansion of emerging energy storage technologies such as vanadium redox flow batteries [4], the global demand for vanadium has exhibited a marked upward trend. Consequently, the development of clean and efficient methods for vanadium extraction from vanadium-bearing shale carries substantial practical and strategic significance.

Oxalic acid is a naturally occurring organic acid with strong reducing and complexing properties, widely employed in hydrometallurgy for the leaching and separation of metal ions, such as in iron recovery from red mud [5,6,7]. It serves as an efficient and environmentally benign leaching agent in the extraction of vanadium from vanadium-bearing shale [8]. Compared to conventional strong acid systems, oxalic acid not only effectively dissolves and liberates vanadium but also significantly mitigates adverse environmental impacts. Studies have demonstrated that oxalic acid enables high-efficiency vanadium leaching from vanadium-bearing shale, whether through direct leaching or roasting-leaching processes. Moreover, iron can be selectively precipitated as ferrous oxalate (FeC₂O₄), facilitating effective separation of vanadium and iron [9,10]. However, due to the isomorphic substitution of vanadium within the muscovite crystal lattice in vanadium-bearing shale [11,12], aluminum ions are concurrently leached as impurities. This leads to a complex oxalic acid leachate containing multiple metallic species, with particularly high concentrations of aluminum, which interferes with the selective recovery of vanadium [13]. Therefore, efficient vanadium separation requires the deep removal of aluminum ions from the leachate.

Solvent extraction is an efficient technique for the selective separation of target metals from complex multielement aqueous solutions [14,15,16]. By selecting appropriate extractants and tuning the conditions of the aqueous phase, efficient separation of the target metal ion—whether present as a cation or anionic complex—from coexisting impurity ions can be achieved [17,18,19]. The oxalic acid leachate derived from vanadium-bearing shale contains multiple high-valent metal ions, including vanadium, iron, and aluminum. These ions exist in the oxalic acid system as stable anionic complexes, such as [VO(C₂O₄)₂]²⁻, [Fe(C₂O₄)₂]⁻, and [Al(C₂O₄)₂]⁻ [20,21,22]. Differences in the stability and coordination behavior of these complexes provide a basis for their selective separation via solvent extraction. Research indicates that the selectivity in vanadium extraction depends critically on the coordination interaction between the functional groups of the extractant and the vanadium complexes, as well as on the competitive partitioning behavior of various metal oxalate complexes between the aqueous and organic phases [23,24]. In our previous study [25], using Aliquat 336 as the extractant under optimized conditions, selective separation of vanadium and aluminum was achieved, with a vanadium extraction efficiency of 99.06% and an aluminum co-extraction rate of only 7.95%. Through analytical techniques such as FTIR and ESI-MS, combined with thermodynamic modeling, it was determined that vanadium and aluminum predominantly exist as [VO(C₂O₄)₂]²⁻ and [Al(C₂O₄)₂]⁻, respectively, in the oxalic acid leachate of vanadium-bearing shale. This confirmed that the extraction mechanism for both metals follows an anion exchange pathway, wherein [VO(C₂O₄)₂]²⁻ forms an ion-association complex with the quaternary ammonium cation (R₄N⁺) and transfers into the organic phase. However, since both [VO(C₂O₄)₂]²⁻ and [Al(C₂O₄)₂]⁻ are anionic species, the molecular-level origin of the selective extraction of vanadium over aluminum by the quaternary ammonium-based extractant Aliquat 336 remains unclear and has not been systematically investigated.

Quantum chemistry, as a high-precision computational tool in materials science, has been widely applied in recent years to chemical systems such as organic molecules, organometallic complexes, and intricate multicomponent media [26]. With regard to the molecular structure and properties of extractants, quantum chemical calculations and geometry optimizations enable the elucidation of electronic structural features and relative stabilities of metal oxalate complexes at the molecular level [27,28], thereby providing a theoretical foundation for understanding the selective separation mechanisms of vanadium and aluminum in quaternary ammonium-based extraction systems. In this study, density functional theory (DFT) calculations are employed. Building upon previous experimental findings, interfacial extraction models involving [VO(C₂O₄)₂]²⁻ and [Al(C₂O₄)₂]⁻ with quaternary ammonium cations (R₄N⁺) are constructed to investigate the mechanism underlying the selective extraction of vanadium over aluminum from oxalic acid leachate of vanadium-bearing shale. This work thus contributes to refining the theoretical framework for purification and enrichment processes within oxalic acid-based hydrometallurgical systems.

2. Experimental Methods and Reagents

2.1. Composition Analysis of Oxalic Acid Leachate of Vanadium-Bearing Shale

The vanadium-bearing shale was sourced from Hubei. Its phase composition is shown in Figure 1 [29], multi-element chemical analysis is shown in

Table 1. Chemical multi-element analysis results of vanadium-bearing shale (wt.%).

Constituent	V2O5	Al2O3	Fe2O3	MgO	CaO	SiO2	K2O	Na2O	S
Content	0.75	9.26	4.23	1.45	3.13	59.45	2.04	0.48	3.72

.

Based on the previously reported methodology [25], the preparation of oxalic acid leachate from vanadium-bearing shale was carried out as follows: the raw shale sample was roasted at 850 °C for 1 h, followed by leaching at 95 °C for 4 h using 60% oxalic acid and 5% CaF₂, with the L/S = 1:1. After leaching, iron ions were selectively removed from the solution to complete the leachate [9,10]. The initial pH of the oxalic acid leachate used in this study was 0.65. The chemical composition of the oxalic acid leachate of shale was analyzed, and the results are presented in

Table 2. Main Chemical Composition of the Oxalic Acid Leachate of shale (g/L).

Constituent	V	Al	K	Na	P	F	C2O42−
Content	1.57	14.23	3.47	0.56	1.22	8.37	120.25

.

The results presented in Table 2 indicated that the vanadium concentration in this pickle liquor was relatively low, at only 1.57 g/L. However, the solution contains a variety of impurity ions at high concentrations. Notably, the aluminum concentration reaches as high as 14.23 g/L. Therefore, the pickle liquor constitutes a complex system characterized by low vanadium content, high acidity, and the presence of multiple interfering species, rendering the purification and enrichment of vanadium from this solution particularly challenging.

2.2. Reagents Used in the Experiments

The Aliquat 336 extractant was provided by Zhengzhou Qinshi Technology Co., Ltd. (Zhengzhou, China), and its structure is shown in Figure 2. The oxalic acid and Tri-butyl phosphate (TBP) were purchased from Sinopharm Chemical Reagent Co., Ltd. (Shanghai, China). All reagents were of analytical grade.

2.3. Extraction Experiments

Organic phase preparation: The extractant (Aliquat 336), phase modifier (TBP), and diluent (sulfonated kerosene) were poured into a beaker in specified proportions, stirred evenly to obtain the organic phase.

Extraction experiment: A magnetic stir bar was placed in a beaker, which was then positioned on a magnetic stirrer. The pickle liquor and organic phase were added to the beaker at the specified volume ratio and mixed under controlled stirring for the designated extraction time. Following the reaction, the biphasic mixture was transferred to a separatory funnel to allow complete phase separation, yielding the loaded organic phase and the raffinate. Assuming negligible volume changes during the extraction process, the extraction efficiency (E₁), distribution ratios (D), separation factor of ions a and b after extraction (β_a,b) were calculated according to Equations (1)–(5), respectively.

$$E_1={\displaystyle \frac{C_0-C_1}{C_0}}\times 100\%$$

(1)

$$D={\displaystyle \frac{C_1}{C_2}}$$

(2)

$$\beta ={\displaystyle \frac{D_V}{D_{Al}}}$$

(3)

In the formula: C₀, C₁, and C₂ are the ion concentrations in the feed solution, raffinate, and loaded organic phase, respectively (g/L), V₁ and V₂ are the volumes of the feed solution and loaded organic phase, respectively (L), D_V and D_Al were distribution ratios of vanadium and aluminum, respectively.

2.4. Determination Method

The grades and concentrations of vanadium were accurately determined using the ammonium ferrous sulfate titration method [30], a well-established technique for redox-based quantification. Oxalate concentration was quantified via potassium permanganate titration [31], ensuring reliable measurement under acidic conditions. The grades and concentrations of other elements in both vanadium-bearing shale and the corresponding leachate were systematically analyzed by inductively coupled plasma optical emission spectrometry (ICP-OES, IRIS Intrepid II, Thermo Electron Corporation, Waltham, MA, USA), providing high sensitivity and multi-element detection capability. Phase compositions were comprehensively characterized by X-ray diffraction (XRD, D8 Advance, Bruker, Berlin, Germany), enabling definitive identification of crystalline phases. Together, this suite of analytical methods ensures a robust and accurate assessment of both elemental distribution and mineralogical characteristics.

2.5. Quantum Chemical Calculations

This section employed Gaussian 16W software c01 to perform structural optimization of the proposed extracted complexes in the vanadium–aluminum oxalic acid system. Frequency calculations were conducted on the optimized geometries to obtain thermodynamic parameters. The reaction Gibbs free energy (∆G) for each extraction process was calculated, enabling a comparative analysis of the binding affinities between the extractant and the vanadium or aluminum complexes.

In quantum chemical calculations using Gaussian 16W, the initial step involved constructing the molecular structural models, followed by analysis of the computational results. Structural optimization was subsequently carried out within the framework of Density Functional Theory (DFT) using the B3LYP/6-31G(d) basis set for the extractant, the vanadium and aluminum oxalate complexes, and their corresponding extracted complexes. Thereafter, frequency calculations were performed on all optimized structures under the same theoretical level to confirm the nature of stationary points (minima) and to derive thermodynamic corrections. The Gibbs free energy changes (∆G) for both extraction and reverse extraction (stripping) reactions were computed and compared, allowing for an assessment of the relative binding energies associated with the extraction and recovery of vanadium and aluminum.

3. Results and Discussion

3.1. Selective Extraction and Separation of Vanadium and Aluminum

Based on previous research results [9,10,25], the optimal conditions for vanadium-aluminum extraction separation were determined. Under these conditions, a six-stage counter-current extraction of vanadium was conducted. Test conditions: initial pH of 0.65, Aliquat 336 concentration of 40%, TBP concentration of 20%, O/A = 1:2, and extracted for 1 min. The results were shown in

Table 3. Results of Six-Stage Counter-Current Extraction.

Items	Concentration (g/L)		Extraction Efficiencies (%)		Separation Factor
	V	Al	V	Al	V/Al
Loaded Organic Phase	3.10	2.42	98.60	8.50	757.14
Raffinate	0.022	13.02	1.40	91.50
Initial Solution	1.57	14.23	100.00	100.00

.

The results presented in Table 3 demonstrated that after a six-stage counter-current extraction of the pickle liquor, a loaded organic phase containing vanadium at a concentration of 3.10 g/L was obtained. The vanadium extraction reached 98.60%, with the vanadium concentration in the raffinate reduced to as low as 0.022 g/L, indicating highly effective enrichment and recovery of vanadium through Aliquat 336 extraction. Moreover, the aluminum co-extraction was only 8.50%, resulting in a separation factor of 757.14 between vanadium and aluminum, thereby achieving efficient and selective separation of vanadium from aluminum.

3.2. Molecular Structure and Reactive Sites of Oxalic Acid Leachate and Aliquat 336

3.2.1. Speciation of Vanadium and Aluminum in the Oxalic Acid Leachate

Vanadium and aluminum can form various complexes with oxalate ions. According to our previous study, the speciation of vanadium and aluminum in the oxalic acid leachate have been studied [25]. The results are shown in Figure 3.

As shown in Figure 3a, when the pH < 0, vanadium existed predominantly as VOC₂O₄ molecules; whereas at pH values > 0, it primarily formed [VO(C₂O₄)₂]²⁻ anionic complexes. At a pH of 0.65, the proportion of [VO(C₂O₄)₂]²⁻ reached 92.65%, indicating that vanadium in the oxalic acid leachate mainly existed in this anionic form. As illustrated in Figure 3b, aluminum in the leachate could form not only oxalate complexes but also various fluoride-containing complexes. The pH range from 0 to 1.5, aluminum predominantly existed as [Al(C₂O₄)₂]⁻ anions. When the pH exceeded 1.5, aluminum stabilized in the form of [Al(C₂O₄)₃]³⁻ anionic complexes. At the experimental pH of 0.65, the relative proportions of [Al(C₂O₄)₂]⁻ and [Al(C₂O₄)₃]³⁻ were 48.95% and 8.06%, respectively. Given that the initial pH of the oxalic acid leachate used in this study was 0.65, vanadium and aluminum primarily occurred as [VO(C₂O₄)₂]²⁻ and [Al(C₂O₄)₂]⁻, respectively. Furthermore, the presence of these complexes in the shale-derived oxalic acid leachate was experimentally confirmed through analytical techniques such as FT-IR and ESI-MS [25]. The molecular structures of the identified complexes are depicted in Figure 4 and Figure 5.

3.2.2. Molecular Structure and Reactive Sites of Aliquat 336

Aliquat 336 is a quaternary ammonium extractant with the simplified molecular formula [R₄N⁺·Cl⁻]. The chloride ion (Cl⁻) in Aliquat 336 is associated with the organic cationic group [R₄N]⁺ through an ionic bond, rather than being covalently bonded. During the extraction process, the mechanism proceeds via an anion exchange reaction, in which Cl⁻ was displaced by the [VO(C₂O₄)₂]²⁻ anion, leading to the formation of the extracted complex [(R₄N)₂·VO(C₂O₄)₂] [25]. Therefore, this study initiated the computational analysis by performing structural optimization of the [R₄N]⁺ functional group in Aliquat 336. The optimized geometry is presented in Figure 6, and its structural parameters are summarized and analyzed in

Table 4. Structural Information of [R₄N]⁺.

Bond Length/Å			Mulliken Charges				ΔG(kJ/mol)
N-C(N-R)	N-C(N-CH3)	C-H(N-CH3)	N	C(N-R)	C(N-CH3)	H(N-CH3)	−273.02 × 104
1.53	1.50	1.09	+0.255	+0.296	−0.330	+0.210

.

Figure 6 illustrated that the [R₄N]⁺ cation consisted of a central nitrogen atom, three octyl chains (C₈H₁₇), and one methyl group. The three alkyl chains and the nitrogen atom were arranged in a nearly planar, symmetric configuration, with the methyl group extending perpendicularly from the center of this plane.

The data presented in Table 4 showed that the C–N bond lengths between the nitrogen atom and the carbon atoms of the three octyl chains were all 1.53 Å, whereas the C–N bond length connecting the methyl group to the nitrogen atom was slightly shorter at 1.50 Å. The Mulliken charge on the central nitrogen atom was +0.255, while that on the carbon atom of the methyl group was −0.330. The three hydrogen atoms of the methyl group each carried a Mulliken charge of +0.210, resulting in a net Mulliken charge of +0.300 for the entire methyl group. Additionally, the Gibbs free energy (∆G) of [R₄N]⁺ was calculated to be −273.02 × 10⁴ kJ/mol based on frequency analysis.

Frontier molecular orbital theory was further employed to investigate the reactive sites on the molecular surface of [R₄N]⁺. According to this theory, positively charged species, which were prone to accept electrons, should be analyzed in terms of their Lowest Unoccupied Molecular Orbital (LUMO) and electrostatic potential distribution. In contrast, negatively charged species, which tended to donate electrons, required analysis of the Highest Occupied Molecular Orbital (HOMO) and electrostatic potential. Since [R₄N]⁺ carries a net positive charge, only LUMO and electrostatic potential analyses were conducted for the organic-phase cationic structure involved in oxalic acid extraction. The results are shown in Figure 7.

The results presented in Figure 7 indicated that the LUMO of [R₄N]⁺ was localized on the methyl group, which appeared enveloped in the orbital distribution, suggesting that this region possesses the lowest energy and was therefore most favorable for electron acceptance. As shown in Figure 7b, the methyl group in [R₄N]⁺ exhibited a deep blue color in the electrostatic potential map, indicating a strong positive potential. This confirmed its electron-acceptor character and high propensity to interact with electron-rich species during chemical reactions.

3.2.3. Molecular Structure and Reactive Sites of VO(C₂O₄)₂²⁻

Vanadium mainly existed in the oxalic acid leachate as the VOC₂O₄ molecule and the VO(C₂O₄)₂²⁻ complex anion. The coordination reactions are shown in Equations (4) and (5) [32].

$${\mathrm{VO}}^{2+}{+\mathrm{C}}_2{{\mathrm{O}}_4}^{2-}{=\mathrm{VOC}}_2{\mathrm{O}}_4$$

(4)

$${\mathrm{VOC}}_2{\mathrm{O}}_4{+\mathrm{C}}_2{{\mathrm{O}}_4}^{2-}=\mathrm{VO}{{({\mathrm{C}}_2{\mathrm{O}}_4)}_2}^{2-}$$

(5)

According to Equations (4) and (5), vanadium-oxalate complexes were formed by the coordination reaction between VO²⁺ and C₂O₄²⁻, producing VOC₂O₄ and VO(C₂O₄)₂²⁻. Therefore, the molecular structures of VO²⁺ and C₂O₄²⁻ were first optimized, as shown in Figure 8.

Figure 8a illustrated that the electrostatic potential of the VO²⁺ ion was deep blue at the vanadium atom, indicating a strong positive potential and pronounced electron-acceptor character. Figure 8b showed that the C₂O₄²⁻ molecule exhibits high structural symmetry, with the electrostatic potential around the oxygen atoms appearing red and uniformly distributed across the four equivalent oxygen sites, reflecting clear electron-donor characteristics.

Based on the above analysis, the structure of the VOC₂O₄ complex formed by VO²⁺ and C₂O₄²⁻ was optimized, as shown in Figure 9a. Subsequently, the [VO(C₂O₄)₂]²⁻ complex was further optimized, and its structural features are presented in Figure 9b. The corresponding structural parameters were analyzed and summarized in

Table 5. Molecular Structure Information of VO(C₂O₄)₂²⁻.

Bond Length/Å				Mulliken Charges					ΔG (kJ/mol)
V=O	V-O	C=O	C-O	V	C	O(V=O)	O(V-O)	O(C=O)
1.60	1.99	1.22	1.30	+1.188	+0.225	−0.560	−0.465	−0.417	−465.72 × 104

.

Based on Figure 9 and Table 5, the [VO(C₂O₄)₂]²⁻ complex exhibited a symmetrical structure centered around the V=O double bond, with two oxalate groups coordinating to the vanadium center in a bidentate fashion. The bond angle between the two oxalate ligands was 142°. The V=O bond length was 1.60 Å, while the V–O bond lengths were 1.99 Å. Within the oxalate groups, the C=O and C–O bond lengths were 1.22 Å and 1.30 Å, respectively. The Mulliken charge on the vanadium atom was +1.188, indicating significant electron deficiency. The oxygen atom in the V=O bond carried a Mulliken charge of −0.560, whereas the oxygen atoms involved in the V–O coordination bonds and the C=O bonds had charges of −0.465 and −0.417, respectively. Compared to the free oxalate ion, the oxygen atoms within the coordinated oxalate groups no longer display uniform charge distribution; specifically, those participating in V–O coordination exhibited higher negative charge density than those in C=O bonds. Additionally, the Gibbs free energy (∆G) of [VO(C₂O₄)₂]²⁻ was calculated to be −465.72 × 10⁴ kJ/mol based on frequency analysis.

Figure 10a illustrated that the HOMO of [VO(C₂O₄)₂]²⁻ was localized on the eight oxygen atoms of the two oxalate ligands, indicating that these atoms possess relatively high orbital energy and were thus prone to electron donation. Notably, the oxygen atoms involved in the V–O single bonds exhibited greater electron-donating tendency compared to those in the C=O double bonds. Figure 10b showed that the entire [VO(C₂O₄)₂]²⁻ complex appeared predominantly red in the electrostatic potential map, with the highest potential localized between the two oxygen atoms in the –C–O–V–O–C– linkage. This distribution confirmed its strong electron-donor character and suggested a high reactivity toward species with low-lying unoccupied orbitals during chemical interactions.

3.2.4. Molecular Structure and Reactive Sites of Al(C₂O₄)₂²⁻

Aluminum mainly existed in the oxalic acid leachate as the Al(C₂O₄)²⁻ complex anion. The coordination reactions are shown in Equations (6) and (7) [33].

$${\mathrm{Al}}^{3+}{+\mathrm{C}}_2{{\mathrm{O}}_4}^{2-}{=\mathrm{AlC}}_2{{\mathrm{O}}_4}^{+}$$

(6)

$${\mathrm{AlC}}_2{{\mathrm{O}}_4}^{+}{+\mathrm{C}}_2{{\mathrm{O}}_4}^{2-}=\mathrm{Al}{{\left({\mathrm{C}}_2{\mathrm{O}}_4\right)}_2}^{-}$$

(7)

According to Equations (6) to (7), Al³⁺ coordinated with C₂O₄²⁻ to form AlC₂O₄⁺ and Al(C₂O₄)²⁻. Then, the structures of AlC₂O₄⁺ and Al(C₂O₄)²⁻ were optimized, as shown in Figure 11. Their structural information is analyzed in

Table 6. Molecular Structure Information of Al(C₂O₄)²⁻.

Bond Length/Å			Mulliken Charges				ΔG (kJ/mol)
Al-O	C=O	C-O	Al	C	O(Al-O)	O(C=O)	Al(C2O4)2−
1.77	1.20458	1.32997	+0.917	+0.279	−0.421	−0.338	−261.80

.

Based on Figure 11 and Table 6, the [Al(C₂O₄)₂]⁻ complex featured a central aluminum atom coordinated by two oxalate ligands in a vertically symmetric, cross-shaped configuration. The Al–O bond length in [Al(C₂O₄)₂]⁻ was 1.77 Å, while the C=O and C–O bond lengths were 1.20 Å and 1.33 Å, respectively. Compared to the V–O bond in [VO(C₂O₄)₂]²⁻ (1.99 Å), the Al–O bond was notably shorter, indicating a stronger interaction between aluminum and the coordinating oxygen atoms. The Mulliken charge on the aluminum atom was +0.917, reflecting significant electron deficiency. The Mulliken charges on the oxygen atoms involved in the Al–O coordination bonds and the C=O bonds were −0.421 and −0.334, respectively. In contrast to the free oxalate ion (C₂O₄²⁻), the four oxygen atoms in the coordinated oxalate groups no longer exhibited uniform charge distribution, with those in Al–O bonds carrying greater negative charge density than those in C=O bonds. Additionally, the Gibbs free energy (∆G) of [Al(C₂O₄)₂]⁻ was calculated to be −261.80 × 10⁴ kJ/mol based on frequency analysis.

To further identify the reactive sites of [Al(C₂O₄)₂]⁻, frontier molecular orbital theory was employed to investigate the distribution of active sites on its molecular surface. The results are presented in Figure 12.

Figure 12a illustrated that the HOMO of [Al(C₂O₄)₂]⁻ was localized on the eight oxygen atoms of the two oxalate ligands, indicating that these atoms possess relatively high energy and were prone to electron donation. Notably, the oxygen atoms in the C=O double bonds exhibited a greater tendency for electron loss compared to those involved in the Al–O coordination bonds. Figure 12b showed that the periphery of the [Al(C₂O₄)₂]⁻ complex appeared predominantly red in the electrostatic potential map, with a deeper red region observed between the C=O double bonds, indicating higher negative charge density at these sites. This distribution confirmed its electron-donor character and suggested a strong propensity to interact with electron-deficient species through orbital overlap or electrostatic attraction during chemical reactions.

3.3. Mechanism of the Selective Separation of Vanadium and Aluminum

3.3.1. Structural Optimization of the VO(C₂O₄)₂²⁻ Extracted Complex Molecule

The structure of [R₄N·VO(C₂O₄)₂]⁻ was optimized, as shown in Figure 13. Subsequently, [(R₄N)₂·VO(C₂O₄)₂] was further structurally optimized, as shown in Figure 14.

Figure 13 illustrated that the extracted complexes [R₄N·VO(C₂O₄)₂]⁻ and [(R₄N)₂·VO(C₂O₄)₂] were formed through interactions between two hydrogen atoms from the methyl group of [R₄N]⁺ and two oxygen atoms from the V–O bonds in [VO(C₂O₄)₂]²⁻. Stable H…O hydrogen bonds were established between these H and O atoms, with bond lengths of 2.24 Å and 2.49 Å, respectively. The electrostatic potential was highest between the oxygen atoms involved in the V–O coordination bonds of [VO(C₂O₄)₂]²⁻, confirming their strong electron-donor character. As a result, the hydrogen atoms of the methyl group in [R₄N]⁺ readily interacted with the electron-rich oxygen atoms in [VO(C₂O₄)₂]²⁻ to form stable complex species. Based on frequency calculations, the Gibbs free energy (∆G) of [R₄N·VO(C₂O₄)₂]⁻ and [(R₄N)₂·VO(C₂O₄)₂] was determined to be −738.76 kJ/mol and −1011.83 kJ/mol, respectively.

3.3.2. Structural Optimization of the Al(C₂O₄)₂⁻ Extracted Complex Molecule

The molecular structure of [R₄N·Al(C₂O₄)₂] was optimized, as shown in Figure 14. Figure 14 illustrated that the extracted complex [R₄N·Al(C₂O₄)₂] was formed through interactions between two hydrogen atoms from the methyl group of [R₄N]⁺ and two oxygen atoms from the C=O double bonds in [Al(C₂O₄)₂]⁻. Stable H…O hydrogen bonds were established between these atoms, with bond lengths of 2.25 Å and 2.71 Å, respectively. The HOMO distribution and electrostatic potential were higher at the oxygen atoms of the C=O double bonds compared to those involved in the Al–O coordination bonds in [Al(C₂O₄)₂]⁻, indicating greater electron density and stronger electron-donor character at these sites. Consequently, the hydrogen atoms of the methyl group in [R₄N]⁺ preferentially interacted with the oxygen atoms of the C=O double bonds in [Al(C₂O₄)₂]⁻ to form stable complex species. Based on frequency calculations, the Gibbs free energy (∆G) of [R₄N·Al(C₂O₄)₂] was determined to be −536.94 kJ/mol.

3.3.3. DFT Calculations Reveal the Separation Mechanism

Based on the above results, energy calculations were performed on the optimized structures of vanadium and aluminum. The extraction ∆G values for Aliquat 336 with vanadium and aluminum were calculated according to Equations (8)–(10), with results shown in

Table 7. Extraction Reaction ∆G for Aliquat 336 with Vanadium and Aluminum.

Extraction Reaction Equation	ΔG (kJ/mol)
[R4N]+ + VO(C2O4)22− = [R4N·VO(C2O4)2]−	−287.96
2[R4N]++VO(C2O4)22− = [(R4N)2·VO(C2O4)2]	−848.21
[R4N]++Al(C2O4)2− = [R4N·Al(C2O4)2]	−186.68

.

$${\left[{\mathrm{R}}_4\mathrm{N}\right]}^{+}+\mathrm{VO}{\left({\mathrm{C}}_2{\mathrm{O}}_4\right)}_2^{2-}={\left[{\mathrm{R}}_4\mathrm{N}\cdot \mathrm{VO}{\left({\mathrm{C}}_2{\mathrm{O}}_4\right)}_2\right]}^{-}$$

(8)

$${2[\mathrm{R}}_4{\mathrm{N}]}^{+}+\mathrm{VO}{\left({\mathrm{C}}_2{\mathrm{O}}_4\right)}_2^{2-}=\left[({\mathrm{R}}_4{\mathrm{N})}_2\cdot \mathrm{VO}{\left({\mathrm{C}}_2{\mathrm{O}}_4\right)}_2\right]$$

(9)

$${\left[{\mathrm{R}}_4\mathrm{N}\right]}^{+}{+\mathrm{Al}(\mathrm{C}}_2{\mathrm{O}}_4{)}_2^{-}=\left[({\mathrm{R}}_4\mathrm{N})\cdot {\mathrm{Al}(\mathrm{C}}_2{\mathrm{O}}_4{)}_2\right]$$

(10)

The Gibbs free energy change (∆G) for the extraction reaction between [VO(C₂O₄)₂]²⁻ and [R₄N]⁺ was −287.96 kJ/mol and −848.21 kJ/mol for the two stoichiometric forms, respectively, while that for [Al(C₂O₄)₂]⁻ with [R₄N]⁺ was −186.68 kJ/mol. These results indicated that aluminum exhibits a lower thermodynamic driving force for binding with Aliquat 336 compared to vanadium species. The binding affinity order of vanadium and aluminum toward Aliquat 336 was: vanadium > aluminum. This demonstrated that during the extraction process, vanadium was preferentially extracted over aluminum. Therefore, in the treatment of pickle liquor, selective separation of vanadium and aluminum can be achieved due to the difference in their binding energies with Aliquat 336. The extraction and separation mechanism for vanadium and aluminum is illustrated in Figure 15:

4. Conclusions

This study elucidated the extraction behavior and separation mechanism of vanadium and aluminum impurities from oxalic acid leachate of vanadium-bearing shale, identified the optimal process conditions and key parameters for efficient extraction, and provides a comprehensive understanding of the selective separation mechanism between vanadium and aluminum.

(1)

Using a mixture of 40% Aliquat 336, 20% TBP, and 40% sulfonated kerosene as the extractant, the pickle liquor was subjected to six-stage counter-current extraction under the following conditions: initial pH of 0.65, O/A = 1:2, and contact time of 1 min. Under these conditions, the extraction efficiencies for vanadium and aluminum reached 98.60% and 8.50%, respectively. The significant difference in extraction behavior enabled effective selective separation of vanadium from aluminum. Compared with the conventional sulfuric acid system, solvent extraction in the oxalic acid system offers several advantages: including faster extraction kinetics, no requirement for initial pH adjustment, and the absence of oxidation or reduction pretreatment steps.

(2)

Molecular structure optimizations were performed using Gaussian 16W software for the extractant Aliquat 336, as well as for the vanadium and aluminum oxalate complexes. Subsequently, the structures of the corresponding extraction complexes formed with vanadium and aluminum were further optimized. Finally, the Gibbs free energy changes (∆G) for the respective extraction reactions were calculated. The results indicated that the binding affinity of Aliquat 336 toward the metal complexes follows the order: vanadium (−287.96 kJ/mol) > aluminum (−186.68 kJ/mol), demonstrating that vanadium is preferentially extracted over aluminum under the given conditions. This thermodynamic preference enables the selective separation of vanadium from aluminum at appropriate extractant concentrations.