Green and economical lithium recovery from spent LiMn₂O₄ batteries through blue vitriol-driven selective sulfation roasting

Abstract

Conventional lithium-ion battery recycling technologies suffer from poor economic performance, secondary pollution, and limited lithium selectivity, especially for low-value LiMn₂O₄ batteries. This study proposes an innovative blue vitriol-driven sulfation roasting–water leaching process for selective lithium extraction from spent LiMn₂O₄ batteries. Thermodynamic evaluations demonstrated the preferential conversion of lithium into water-soluble Li₂SO₄ while manganese and copper remain as insoluble oxides. Thermogravimetric analysis identified an optimal temperature range for efficient blue vitriol decomposition and selective lithium conversion. Under optimized conditions (700 °C, 0.7 : 1, 60 min), the sulfation roasting process exhibited 95.18% lithium recovery efficiency and 99.66% lithium selectivity while achieving near-zero SO₂ emissions. Comprehensive characterization elucidated the transformation pathways from LiMn₂O₄ to Li₂SO₄, CuMn₂O₄, and Mn₂O₃. Density functional theory calculations revealed the preferential replacement of lithium in LiMn₂O₄ by copper due to stronger Cu–O bonding and higher electron delocalization. The recovered Li₂CO₃ attained 99.59% purity, satisfying commercial battery-grade standards. Economic analysis revealed a net profit of $1 689.31 per metric ton of spent LiMn₂O₄ batteries, with additional revenue potential from byproducts (e.g. CuMn₂O₄ spinel). This research establishes an environmentally benign and economically viable approach for selective lithium extraction from low-value lithium-ion batteries, while providing a framework for broader battery recycling applications.

---

Green foundation

1. This study presents an innovative sulfation roasting process utilizing blue vitriol for the sustainable recycling of spent LiMn₂O₄ batteries. This research overcomes the limitations of conventional recycling methods by eliminating the reliance on high temperatures (>1000 °C) and excessive chemical reagents, while bypassing complex metal separation steps.

2. The process achieves high lithium recovery efficiency (95.18%) and lithium selectivity (99.66%) under optimized conditions, while the obtained Li₂CO₃ product meets battery-grade standards (99.59% purity). Crucially, the entire roasting process achieves near-zero emissions of SO₂, thereby eliminating secondary pollution risks.

3. The water-leached residue after Li recovery consists of CuMn₂O₄, CuO, and Mn₂O₃, and has excellent redox properties. The future work can investigate its catalytic performance for the removal of pollutants such as VOCs and CO.

1. Introduction

Lithium-ion batteries (LIBs) have become the preferred choice for electric vehicles and smartphones due to their superior electrochemical performance, high energy density, and absence of the memory effect.^1–5 Industry forecasts indicate that global LIB production will exceed 1.3 TWh by 2030.⁶ However, with a limited lifespan of 5–10 years, the world is expected to face a massive wave of retired LIBs in the near future.^3,7,8 Spent LIBs contain metallic components and organic electrolytes.⁹ Improper handling will lead to metal dissolution and electrolyte leakage, causing severe environmental pollution.^10–12 Moreover, the contents of valuable metals (e.g., Li, Ni, Co, and Mn) in spent LIBs are far higher compared to natural ores, making them an important secondary resource.^13,14 Therefore, developing efficient recycling methods for valuable metals from spent LIBs is crucial for both environmental protection and sustainable resource utilization.

The mainstream approaches for recovering valuable metals from spent LIBs include pyrometallurgy and hydrometallurgy.^3,15 Pyrometallurgy employs high-temperature smelting to reduce valuable metals into metal alloys.¹⁶ This mature approach offers operational simplicity and suitability for large-scale application.^17,18 However, it requires extremely high temperatures of 1000 to 1400 °C, resulting in harmful gas emissions and massive lithium loss through volatilization.^19,20 Hydrometallurgy dissolves battery materials using acid or alkali (e.g., HCl and NH₃·H₂O) to leach valuable metals, followed by stepwise separation and recovery of various valuable metals.²¹ Compared with pyrometallurgy, hydrometallurgy demonstrates lower operating temperature and a higher metal recovery rate.²² However, it consumes substantial chemical reagents and generates acidic/alkaline wastewater along with harmful gases during leaching.^23,24 It should be noted that these two approaches primarily derive profits from recovering high-value metals (e.g., Co and Li), making them economically viable for high-value LIBs (e.g., LiCoO₂ and LiNi_xCo_yMn_zO₂ batteries). In contrast, these two approaches face severe economic challenges in recycling low-value LIBs (e.g., LiMn₂O₄ batteries).¹¹ The relatively low market value of manganese-based materials often leads to marginal profits or even economic losses.²⁵ Thus, developing cost-effective recycling approaches for comprehensive recovery of various types of spent LIBs has become imperative.

To overcome the limitations of traditional technologies, integrated pyro-hydrometallurgical approaches such as sulfation roasting have gained attention. This approach processes spent LIBs with sulfation agents at 350 °C–800 °C, converting valuable metals into water-soluble sulfates, followed by water leaching.¹⁸ Teng et al. employed NH₄HSO₄ as a sulfation agent to recover valuable metals from ternary lithium-ion batteries. The results revealed that sulfation roasting at 400 °C could achieve >90% recovery for Li, Co, Mn, and Ni.²⁶ Unlike traditional technologies, sulfation roasting–water leaching eliminates the need for high temperatures of 1000 to 1400 °C and acid/alkali consumption while maintaining a high metal recovery rate. However, current sulfation roasting approaches still face several limitations: (1) the application of sulfation agents (e.g., NH₄HSO₄ and CaSO₄·H₂O) increases operational costs; (2) several harmful gases (SO₂ and NH₃) released during roasting cause secondary pollution; (3) impurity cations (e.g., Ca²⁺) from sulfation agents exist in the leachate, reducing the purity of recovered valuable metals; and (4) the non-selective leaching of various valuable metals exhibits a challenge for subsequent separation. As lithium recovery is typically positioned at the terminal stage of the process, this approach inevitably leads to lithium loss. As a result, optimizing the current sulfation roasting approaches is imperative to reduce costs, minimize environmental impact, eliminate impurity contamination, and achieve selective lithium extraction.

Blue vitriol (CuSO₄·5H₂O) is a naturally occurring sulfate mineral with abundant reserves.²⁷ During electrolytic copper refining, large quantities of copper-bearing sludge are generated in electrolyte purification systems. This sludge can be directly processed through evaporation–crystallization to produce high-purity blue vitriol (≥99.5%).²⁸ Thermal decomposition of blue vitriol will generate SO₂ gas and CuO. The evolved SO₂ can subsequently sulfate valuable metals from spent LIBs and convert them into water-soluble metal sulfates. He et al. employed pyrite (FeS₂) for sulfation roasting of LiCoO₂ batteries, verifying the excellent reactivity of SO₂ as a sulfation agent.²⁹ In addition, studies have shown that spinel-type LiMn₂O₄ features a three-dimensional tunnel structure favorable for Li⁺ migration, where the Li–O bond exhibits far weaker binding energy compared to Mn–O.^30,31 This testifies preferential lithium reactivity over manganese, providing feasibility for selective lithium extraction from spent LiMn₂O₄ batteries.³² Based on this, we hypothesize that precise control of blue vitriol dosage can generate an optimal amount of SO₂ to selectively sulfate lithium into water-soluble Li₂SO₄, while Mn and Cu exist in the form of insoluble oxides. This strategy can enable selective lithium extraction and ensure complete sulfur conversion into Li₂SO₄, preventing SO₂ emission. However, to date, few studies have attempted to validate this speculation.

This study proposes a selective lithium extraction strategy for spent LiMn₂O₄ batteries using blue vitriol as a sulfation agent. First, thermodynamic calculations were conducted to analyze the feasibility of sulfation roasting and the transformation behaviors of valuable metals during roasting. Second, thermogravimetry–differential scanning calorimetry was performed to investigate the roasting behaviors of LiMn₂O₄, blue vitriol, and their mixture, providing guidance for selecting appropriate roasting parameters. Third, the effects of roasting temperature, blue vitriol/LiMn₂O₄ mass ratio, and holding time on the recovery efficiencies of Li/Mn/Cu and selective extraction efficiency of Li were systematically explored. Fourth, the physical and chemical properties of samples before and after roasting were comprehensively characterized. Combined with density functional theory calculations, the mechanism responsible for selective lithium extraction was revealed. Fifth, the purity of the obtained Li₂CO₃ product was tested. Finally, the economic and environmental assessments of the sulfation roasting process were conducted to evaluate its industrial application prospects.

2. Materials and methods

2.1. Materials and agents

The spent LiMn₂O₄ (LMO) batteries used in this study were purchased from a battery recycling center in Qingdao, China. The batteries were first immersed in a 10 wt% NaCl solution for 24 h to completely discharge residual electricity.³³ Then, manual disassembly was performed to separate the casing, cathode sheets, anode sheets, and plastic separators. The obtained cathode sheets were calcined in a muffle furnace with a heating rate of 10 °C min⁻¹ to 450 °C, followed by isothermal treatment for 180 min to remove the PVDF binder.³⁴ Finally, the calcined material was crushed and sieved through a 100-mesh screen to separate the active material from aluminum foil. In this case, black LiMn₂O₄ powder was obtained for subsequent experiments.

Elemental analysis indicated that the carbon content of LMO powder was zero, confirming the complete removal of the PVDF binder during calcination. X-ray diffraction (XRD) patterns (Fig. S1) indicate that the LMO powder consisted of LiMn₂O₄ and Mn₃O₄. The formation of Mn₃O₄ was attributed to lithium de-intercalation during the cyclic charging and discharging process. X-ray fluorescence (XRF) results (Table S1) demonstrate a high Mn content of 97.39% and a negligible Al content of 0.48%, confirming the effective separation of LMO powder from the aluminum foil current collector. The LMO powder was digested using aqua regia (HCl : HNO₃ = 3 : 1, v/v), and the concentrations of metal ions in solution were measured by inductively coupled plasma mass spectrometry (ICP-MS). As shown in Table S2, the mass fractions of Li and Mn in LMO powder were 3.72 wt% and 57.14 wt%, respectively. This was consistent with the characteristics of LMO batteries.

2.2. Sulfation roasting–water leaching experiments

Fig. S2 shows the process flow diagram of selective lithium extraction from spent LMO batteries using blue vitriol as a sulfation agent. First, blue vitriol and LMO were uniformly mixed at mass ratios of 0.35 : 1, 0.7 : 1, 1.05 : 1, and 1.4 : 1. The mixtures were then transferred into a rectangular ceramic boat and calcined in a muffle furnace under an air atmosphere. The target temperatures were set as 500 °C, 600 °C, 700 °C, 800 °C, and 900 °C. The holding times were set as 0 min, 30 min, 60 min, 90 min, and 120 min.

After sulfation roasting, the obtained samples were mixed with deionized water at a solid–liquid ratio of 50 g L⁻¹, followed by magnetic stirring at 80 °C for 60 min. Then, the suspension was filtered and repeatedly washed with deionized water to obtain the lithium-rich leachate and solid residue. In this study, the recovery efficiency of Li/Mn/Cu (L_i) and the selective recovery ratio of lithium (S_Li) were calculated according to eqn (1) and (2), respectively.³⁵

where V (L) denotes the volume of the leachate; C_i (g L⁻¹) denotes the element i concentration; m_i (g) denotes the amount of LMO or blue vitriol; and W_i (%) denotes the mass fraction of element i in the raw material; M_i denotes the molar mass of element i.

2.3. Recovery of Li₂CO₃

The lithium-rich leachate was concentrated by heating and stirring in a 95 °C water bath. Then, a saturated Na₂CO₃ solution was slowly added to the concentrate at a molar ratio of Li⁺ : CO₃²⁻ = 1 : 1.2 until the pH value of the solution reached 11. The resulting precipitate was aged for 60 min in a 95 °C water bath to ensure complete crystallization. After that, the solution was thermally filtered at 80 °C. The obtained precipitate was repeatedly washed with hot water (>80 °C) to effectively remove impurity ions (e.g., Na⁺) adsorbed on the surface. Finally, the precipitate was dried in an oven at 80 °C for 8 h, yielding the Li₂CO₃ product.

2.4. Sample characterization

A variety of characterization methods were used to analyze the physical and chemical properties of various samples, including thermal properties, crystal structure, surface micromorphology, chemical valence, and ion content. More information is provided in Text S1 in the SI.

2.5. Density functional theory (DFT) calculations

DFT calculations were performed with the first-principles simulation using the Cambridge Sequential Total Energy Package (CASTEP) module in Materials Studio software. The exchange–correlation potential was described using the generalized gradient approximation (GGA) with the Perdew–Burke–Ernzerhof (PBE) functional. The interactions between valence electrons and ionic cores were described using the OTFG ultrasoft pseudo-potential method. A plane-wave basis set with a cutoff energy of 380 eV was assigned to the potential method. The empirical dispersion correction in Grimme's scheme was employed to consider the van der Waals (vdW) interaction. The Broyden–Fletcher–Goldfarb–Shannon (BFGS) algorithm with a medium quality setting of k-points was used for all the energy minimizations in this work. The geometry optimization convergence tolerances for the energy change, maximum force and maximum displacement were 5 × 10⁻⁵ eV per atom, 0.001 eV Å⁻¹, and 0.005 Å, respectively. For all the models, a 15 Å vacuum space was set in the z-axis to guarantee full relaxation.

3. Results and discussion

3.1. Thermodynamic analysis

Table 1 lists the possible reactions between blue vitriol and LMO under different sulfur input conditions. The standard Gibbs free energy changes (ΔG^⊖) for these reactions in the temperature range of 0–1000 °C were calculated using HSC Chemistry 6.0 software (Fig. 1(a)). Due to the lack of thermodynamic data for LMO, a model system composed of Li₂O, Mn₂O₃, and MnO₂ in specific proportions was used as a substitute.³⁶ At low sulfur addition, blue vitriol selectively sulfated lithium into water-soluble Li₂SO₄, while manganese and copper existed as water-insoluble Mn₂O₃ and CuO (Reaction (1)). At moderate sulfur addition, complete sulfation of lithium occurred along with partial sulfation of manganese into water-soluble MnSO₄ (Reaction (2)). At high sulfur addition, both lithium and manganese underwent complete sulfation, yielding Li₂SO₄ and MnSO₄ (Reaction (3)). As shown in Fig. 1(a), the ΔG^⊖ values of all three reactions became negative when the temperature exceeded 100 °C. This confirmed that the chemical reactions for LMO recovery using blue vitriol were thermodynamically spontaneous. Furthermore, the ΔG^⊖ values decreased sharply with the increase of temperature, indicating that higher temperatures promoted the sulfation process.

Fig. 1 (a) Standard Gibbs free energy changes (ΔG^⊖) of reactions listed in Table 1 in the temperature range of 0–1000 °C. The predominance area diagrams of (b) Cu, (c) Li, and (d) Mn at varying temperatures and SO₂ partial pressures (log p_SO2).

Table 1 The reaction equations of possible reactions

Reaction equations
2CuSO4·5H2O + 4LiMn2O4 → 2Li2SO4 + 2CuO + 4Mn2O3 + 10H2O + O2 (g)	Low sulfur	(R1)
6CuSO4·5H2O + 4LiMn2O4 → 2Li2SO4 + 4MnSO4 + 2Mn2O3 + 6CuO + 30H2O + 2O2 (g)	Moderate sulfur	(R2)
10CuSO4·5H2O + 4LiMn2O4 → 2Li2SO4 + 8MnSO4 + 10CuO + 50H2O + 3O2 (g)	High sulfur	(R3)

---

Fig. 1(b–d) present the predominance area diagrams for Cu, Li, and Mn at varying temperatures and SO₂ partial pressures (log p_SO2). Similar transformation trends were found for all three metals: higher sulfur addition promoted sulfate formation while higher temperatures favored oxide generation. However, obvious differences existed in their specific phase transition conditions. At equivalent SO₂ partial pressure, CuSO₄ decomposed most readily into CuO and SO₂, which then reacted with LMO. Meanwhile, Li₂SO₄ exhibited superior thermal stability with a decomposition temperature far higher than that of MnSO₄ and CuSO₄. This thermal gradient enabled selective lithium extraction – controlling the temperature could decompose CuSO₄/MnSO₄ into water-insoluble oxides while maintaining lithium as water-soluble Li₂SO₄. Moreover, lithium required the lowest SO₂ partial pressure for sulfation at any given temperature, indicating its preferential sulfation over manganese and copper. These phenomena suggested that precise regulation of the roasting temperature and SO₂ partial pressures could realize selective lithium extraction from spent LMO.

3.2. TG-DSC analysis

Fig. 2(a) presents the TG-DSC curves of LMO. A slight mass loss (<3%) appeared at 894–1000 °C, along with an endothermic peak at 941 °C. This corresponded to LMO decomposition, while the minimal mass loss implied the excellent thermal stability of LMO.³⁷Fig. 2(b) displays the TG-DSC curves of blue vitriol. It was found that an obvious weight loss and endothermic peaks appeared at 25–270 °C, attributed to the evaporation of crystal water from blue vitriol. The endothermic peaks at 92 °C and 120 °C were assigned to the release of four water molecules, while the peak at 236 °C represented the release of the final water molecule.²⁷ As the temperature further increased, two distinct endothermic peaks appeared at 745 °C and 793 °C in the DSC curve. This indicated the thermal decomposition of CuSO₄ into CuO and SO₂.³⁸ Meanwhile, the TG curve exhibited massive mass loss due to SO₂ release from CuSO₄ decomposition. It should be noted that CuSO₄ decomposition began at 650 °C, consistent with the phase transition temperature from CuSO₄ to CuO in the Cu predominance diagrams (Fig. 1(b)).

Fig. 2 The TG-DSC curves of (a) LMO, (b) blue vitriol, and (c) LMO–blue vitriol mixture at a mass ratio of 0.8 : 1. (d) Effect of roasting temperature, (e) blue vitriol/LMO mass ratios and (f) holding time on the recovery efficiency of Li/Mn/Cu and Li selective recovery ratio.

Fig. 2(c) illustrates the TG-DSC curves of the LMO–blue vitriol mixture. Similar to blue vitriol, the mixture displayed weight loss and endothermic peaks at 25–270 °C due to the removal of crystalline water. When the temperature increased to 520 °C, the mixture began to exhibit mass loss and exothermic behavior. However, neither pure LMO nor blue vitriol alone showed any reaction at this temperature (Fig. 2(a and b)). This indicated an interaction between LMO and blue vitriol during roasting, which promoted the sulfation reaction. As the temperature further increased, multiple exothermic peaks appeared in the DSC curve. This suggested that the sulfation reaction between blue vitriol and LMO was exothermic. Meanwhile, the TG curve presented a gradually declining trend. This was likely due to the release of O₂ during the sulfation reaction (Reactions (1)–(3)). The mixture stopped releasing heat at 915 °C, which indicated that the sulfation reaction was nearly complete. At 915–981 °C, an obvious mass loss occurred accompanied by an endothermic peak at 947 °C. This peak corresponded to MnSO₄ decomposition and SO₂ release.³⁹ Based on HSC calculations and thermogravimetric analysis, the optimal temperature for the roasting of LMO–blue vitriol mixture was in the range of 520–910 °C.

3.3. Metal recovery efficiency and lithium selectivity

Fig. 2(d) shows the effect of roasting temperature on the Li/Mn/Cu recovery efficiency and Li selective recovery ratio. At 500 °C, L_Cu reached 82.10% while L_Li was only 28.51% with poor Li selectivity (S_Li = 31.67%). This indicated that only partial decomposition of CuSO₄ occurred, generating insufficient SO₂ for effective Li recovery. At this temperature, the reaction between LMO and blue vitriol proceeded through solid-phase ion exchange without SO₂ release or heterogeneous gas–solid reactions. These results were consistent with thermogravimetric analysis, which showed that CuSO₄ decomposition happened above 700 °C (Fig. 2(b)) and sulfation reaction started at 520 °C (Fig. 2(c)). When the temperature increased to 700 °C, L_Cu sharply decreased to 0.01% while L_Li obviously increased to 94.75% with excellent Li selectivity (S_Li = 99.66%). This could be attributed to the complete decomposition of CuSO₄ and subsequent SO₂ release, which facilitated the sulfation reaction. The nearly negligible Mn recovery (L_Mn < 0.26%) demonstrated that Li underwent sulfation reaction preferentially over Mn, consistent with the predominance area diagrams obtained from HSC calculations (Fig. 1(c and d)). However, L_Li dropped to 80.50% when the temperature reached 900 °C. This was likely because of sintering effects that encapsulated Li within dense phases and hindered complete recovery.⁴⁰

Fig. 2(e) shows the effect of blue vitriol/LMO mass ratio on Li/Mn/Cu recovery efficiency and Li selective recovery ratio. When the blue vitriol/LMO mass ratio was 0 : 1 (no blue vitriol addition), both L_Li and L_Mn were low (<6%). This indicated that roasting pure LMO could not effectively destroy its structure. As the blue vitriol/LMO mass ratio increased to 0.35 : 1, L_Li increased rapidly while L_Mn and L_Cu remained nearly zero. This confirmed that Li underwent sulfation preferentially over Mn. At a mass ratio of 0.7 : 1, L_Li and S_Li were 94.75% and 99.66%, respectively. This suggested that highly efficient and selective Li recovery was achieved under these conditions. It should be noted that the S/Li molar ratio in the mixture was 1 : 2 when the mass ratio was 0.7 : 1, which corresponded to Reaction (1) (in Table 1) where only Li was sulfated. At a mass ratio of 0.8 : 1, L_Li reached 100% while L_Mn increased to 4.03%. This indicated that Mn began to sulfate after complete Li sulfation. This observation confirmed the hypothesis from thermogravimetric analysis that MnSO₄ was formed during sulfation roasting (Fig. 2(c)). As the mass ratio further increased to 1.05 : 1, L_Li remained stable at 100% while L_Mn continued to increase. In particular, when the mass ratio was 1.4 : 1, L_Mn increased to 17.09% and S_Li dropped to 75.30%. Such a phenomenon revealed that excess sulfur addition (S/Li molar ratio exceeded 1 : 2) would lead to increasing conversion of Mn into water-soluble MnSO₄, thereby reducing the Li selective recovery ratio. This corresponded to Reactions (2) and (3) (in Table 1) where both Li and Mn were sulfated.

Fig. 2(f) shows the effect of holding time on the Li/Mn/Cu recovery efficiency and Li selective recovery ratio. Compared with roasting temperature and blue vitriol/LMO mass ratio, holding time exhibited a relatively minor effect. When the holding time was extended from 0 to 30 min, L_Li increased from 94.67% to 95.85%. The reason was that prolonged roasting facilitated further conversion of Li into Li₂SO₄. However, a further increase in the holding time did not obviously change L_Li. This indicated that the sulfation reaction had reached equilibrium. When the holding time was extended from 0 min to 60 min, S_Li slightly increased from 98.89% to 99.66%. This could be attributed to the more complete decomposition of CuSO₄, preventing Cu²⁺ interference with Li purity. However, S_Li remained stable when the holding time exceeded 60 min. It was because CuSO₄ had completely decomposed into CuO while Mn remained as metal oxides. Considering L_Li, S_Li, and energy consumption, 60 min was determined as the optimal holding time.

3.4. SO₂ emission

Fig. 3(a) shows the SO₂ emission profiles during the roasting of pure blue vitriol and blue vitriol/LMO mixtures (mass ratios of 0.7 : 1 and 1.4 : 1). All samples were heated at a constant rate of 10 °C min⁻¹ to 700 °C, followed by an isothermal holding period of 30 minutes. For the sample of pure blue vitriol, massive SO₂ emission occurred near 600 °C and reached a maximum of 1286.3 ppm at 680 °C. This was due to the thermal decomposition of CuSO₄.³⁸ Upon entering the holding stage, the SO₂ concentration gradually decreased and eventually dropped below 100 ppm. This demonstrated near-complete decomposition of CuSO₄. At a mass ratio of 0.7 : 1 (S/Li molar ratio of 1 : 2), the maximum SO₂ concentration was only 10.4 ppm. This confirmed that nearly all SO₂ released from CuSO₄ decomposition reacted with LMO to form Li₂SO₄ (Reaction (1)). When the mass ratio increased to 1.4 : 1, the maximum SO₂ concentration reached up to 77.4 ppm. The reason was that when the S/Li molar ratio exceeded 1 : 2, the excess SO₂ would react with Mn to form MnSO₄, which then underwent thermal decomposition to release SO₂.³⁹ These results suggested that precise regulation of blue vitriol dosage not only enabled efficient and selective lithium recovery but also achieved near-zero SO₂ emissions, thereby improving the environmental friendliness of the sulfation roasting process.

Fig. 3 (a) SO₂ emission profiles during the roasting of various samples. XRD patterns of roasted samples under different (b) roasting temperatures, (c) blue vitriol/LMO mass ratios, and (d) holding times.

3.5. Mechanism responsible for selective lithium extraction

Fig. 3(b) exhibits the XRD patterns of roasted samples obtained at different temperatures. When the temperature was 500 °C, the roasted sample mainly consisted of LiMn₂O₄, CuSO₄, and Mn₂O₃. The presence of Mn₂O₃ was due to the oxidation of Mn₃O₄ in the spent LMO battery. This confirmed the weak reaction between blue vitriol and LMO at 500 °C. Although there were water-soluble Li/Cu species in the roasted sample (Fig. 2(d)), their contents were low and thus no corresponding diffraction peaks were detected. When the temperature increased to 600 °C, LiMn₂O₄ peaks weakened and CuSO₄ peaks disappeared. This testified that the reaction between LMO and blue vitriol occurred during roasting. For this reason, new Li₂SO₄ peaks emerged while Mn and Cu existed as Mn₂O₃ and CuO, respectively. This explained the sharp increase of L_Li and the near-zero recovery efficiency of Cu and Mn at 600 °C (Fig. 2(d)). At 700 °C, the LiMn₂O₄ peaks vanished and all Li was transformed into Li₂SO₄. Meanwhile, Mn and Cu remained as metal oxides. In this case, L_Li reached near 100% and the Cu/Mn recovery efficiency was near zero (Fig. 2(d)). Notably, obvious CuMn₂O₄ peaks appeared in the roasted sample. This indicated that higher temperatures promoted the formation of Cu–Mn spinel. When the temperature further increased to 800–900 °C, the intensity of Li₂SO₄ peaks weakened. A possible reason was that partial Li was trapped in dense phases and could not be detected. This explained the sharp decrease of L_Li in the temperature range of 800–900 °C (Fig. 2(d)).

Fig. 3(c) shows the XRD patterns of roasted samples under different blue vitriol/LMO mass ratios. Compared with the fresh LMO sample (Fig. S1), the LiMn₂O₄ peaks still existed and only Mn₃O₄ was replaced by Mn₂O₃. This confirmed that roasting LMO alone could not effectively destroy its structure. When the blue vitriol/LMO mass ratio was 0.35 : 1, the LiMn₂O₄ peaks weakened but remained visible. Meanwhile, the peaks of Li₂SO₄, Mn₂O₃, and CuO appeared. This indicated that the sulfation reaction began to happen at this time (Fig. 2(e)). At a mass ratio of 0.7 : 1, the LiMn₂O₄ peaks disappeared completely. This revealed that all LiMn₂O₄ participated in the sulfation reaction. The Li₂SO₄ peaks also became stronger, and Mn existed in the forms of Mn₂O₃ and CuMn₂O₄. This explained why both L_Li and S_Li exceeded 95% under this condition (Fig. 2(e)). With the further increase in the mass ratio, the Li₂Mn₂(SO₄)₃ peaks appeared and gradually intensified. In particular, when the mass ratio was 1.4 : 1, Li₂SO₄ was completely replaced by Li₂Mn₂(SO₄)₃ while the Mn₂O₃ peaks became weak. The results confirmed that excess blue vitriol would lead to the sulfation of Mn. This explained the phenomenon that L_Mn gradually increased and S_Li gradually decreased when the mass ratio exceeded 0.7 : 1 (Fig. 2(e)).

Fig. 3(d) shows the XRD patterns of roasted samples under different holding times. Comparative analysis of XRD patterns (0–120 min) revealed no obvious changes in the phase composition of various samples. This indicated that holding time had negligible effects on the crystal structure of roasted samples. This observation explained the minimal variation in the values of L_Li/L_Mn/L_Cu with an extended holding time (Fig. 2(e)). However, when the holding time increased from 0 to 30 min, the peak intensity of Li₂SO₄ strengthened. This suggested that a prolonged holding time promoted the formation of Li₂SO₄. This also explained the fact that L_Li increased with the increase in the holding time from 0 to 30 min (Fig. 2(e)). Besides, as the holding time increased, the peak intensity of CuMn₂O₄ gradually intensified. This verified that a prolonged roasting time accelerated the formation of the spinel phase, which would not affect the recovery efficiency of Mn and Cu.

X-ray photoelectron spectroscopy (XPS) was employed to further analyze the valence state changes of Mn and S in the roasted samples. Fig. 4(a–c) show the Mn 2p spectra of various samples. The Mn 2p spectra contained two obvious peaks representing spin–orbit splitting of Mn 2p_3/2 and Mn 2p_1/2.⁴¹ In the absence of blue vitriol, two peaks at 641.4 eV (Mn³⁺) and 642.5 eV (Mn⁴⁺) appeared in the XPS spectra (Fig. 4(a)).^42,43 This verified the co-existence of Mn³⁺ and Mn⁴⁺ in the roasted sample. At a mass ratio of 0.7 : 1, the Mn⁴⁺ peak disappeared and only the peak at 641.4 eV (Mn³⁺) remained in the XPS spectra. This indicated that blue vitriol could reduce Mn⁴⁺ into Mn³⁺ during roasting. Combined with XRD results (Fig. 3(c)), Mn was found to exist in the form of water-insoluble compounds (Mn₂O₃ and CuMn₂O₄) in the roasted sample, thereby enabling selective lithium recovery. As the mass ratio increased to 1.4 : 1, a new peak at 640.4 eV (Mn²⁺) was observed.^44,45 This suggested that excess blue vitriol further reduced Mn³⁺ into Mn²⁺, which was consistent with XRD results that partial Mn₂O₃ was transformed into Li₂Mn₂(SO₄)₃. These findings provided a clear explanation for the phenomenon that L_Mn increased with the increase in blue vitriol dosage.

Fig. 4 XPS spectra of Mn 2p on the (a) LMO and mixtures of blue vitriol/LMO at mass ratios of (b) 0.7 : 1 and (c) 1.4 : 1. XPS spectra of S 2p on the (d) LMO and mixtures of blue vitriol/LMO at mass ratios of (e) 0.7 : 1 and (f) 1.4 : 1.

Fig. 4(d–f) show the S 2p spectra of various roasted samples. As shown in Fig. 4(d), the S 2p spectra of LMO exhibited a higher signal-to-noise ratio compared to the blue vitriol/LMO mixture. This observation confirmed the lower sulfur content in LMO relative to the mixture.^46,47 As shown in Fig. 4(e), at a mass ratio of 0.7 : 1, only one peak assigned to Li₂SO₄ appeared at 168.5 eV.⁴⁸ This revealed that all sulfur in blue vitriol was converted into Li₂SO₄ and existed in the roasted sample, consistent with XRD results (Fig. 3(c)). As shown in Fig. 4(f), at a mass ratio of 1.4 : 1, a new peak assigned to MnSO₄ appeared at 169.7 eV. This demonstrated that excess sulfur would react with Mn to form water-soluble MnSO₄,^49,50 thereby reducing the Li selective recovery ratio (Fig. 2(e)).

Fig. 5(a) and (b) show the microscopic morphology of spent LMO at different magnifications. As shown in Fig. 8a, the spent LMO particles exhibited irregular polyhedral structures with a broad particle size distribution (3–15 μm). To examine the surface characteristics in detail, a representative particle was selected and observed at 20 000× magnification (Fig. 5(b)). The high-resolution image indicated that the spent LMO particles possessed a relatively rough surface with a loose and porous structure. This morphological feature significantly differed from the smooth and dense spinel structure of pristine LiMn₂O₄.⁵¹ The difference could be attributed to mechanical damage caused by lithium intercalation/deintercalation cycles, as well as corrosion effects from electrolyte components on the cathode material.⁵²

Fig. 5 Microscopic morphology of spent LMO at (a) 5000× magnification and (b) 20 000× magnification. Microscopic morphology of the roasted LMO–blue vitriol mixture at (c) 10 000× magnification and (d) 20 000× magnification. Elemental mapping of (e) O element, (f) S element, (g) Mn element, and (h) Cu element on the roasted LMO–blue vitriol mixture.

Fig. 5(c) and (d) show the microscopic morphology of the roasted sample from the LMO–blue vitriol mixture. At 10 000× magnification, the roasted sample exhibited large amounts of spinel-phase grains. These individual spinel grains were relatively small, with particle sizes ranging from 1 μm to 5 μm. Based on the XRD results (Fig. 3(c)), these structures corresponded to spinel-type CuMn₂O₄ generated during sulfation roasting.⁵³ When the magnification was increased to 20 000×, small spherical particles were clearly observed between and around the spinel grains, which were due to the presence of Mn₂O₃ and CuO.^54,55Fig. 5(e–h) show the elemental mapping results of the roasted sample. It was found that the distribution regions of Mn/Cu exhibited higher overlap with O while lower overlap with S. This further confirmed that Mn and Cu mainly existed as their corresponding metal oxides (CuMn₂O₄, Mn₂O₃, and CuO) rather than as sulfides or sulfates in the roasted samples.

Density functional theory calculations were used to reveal the reaction mechanism of selective lithium extraction at the molecular scale. As shown in Fig. S3, initial crystal models of reactants (CuSO₄ and LiMn₂O₄) and possible products (Li₂SO₄, CuMn₂O₄, and Li₂CuO₂) were built and optimized based on the Materials Project database. Among them, Li₂SO₄ was the sulfation product of Li, while CuMn₂O₄ and Li₂CuO₂ were the products of Cu replacing Li and Mn in LiMn₂O₄, respectively. Fig. 6b shows the integrated density of states (DOS) near the Fermi level for these models. The integrated DOS values followed the order CuMn₂O₄ > CuSO₄ > Li₂CuO₂, with Li₂SO₄ having a higher value than LiMn₂O₄. It is well-known that a larger integrated DOS indicates higher electron probability and stronger conductivity.⁵⁶ Therefore, in the reaction between orthorhombic CuSO₄ and spinel LiMn₂O₄, Cu preferentially replaced Li in LiMn₂O₄, forming CuMn₂O₄ and Li₂SO₄. In contrast, the pathway where Cu replaced Mn to form Li₂CuO₂ was energetically unfavorable. The projected densities of states (PDOS) of CuMn₂O₄, LiMn₂O₄ and Li₂CuO₂ were analyzed (Fig. 6(b–d)). In LiMn₂O₄, the PDOS contribution of Li near the Fermi level was much lower than that of Mn. Similarly, in Li₂CuO₂, Li's contribution was significantly lower than Cu's. This indicated weaker Li–O bonding, making Li⁺ highly active and easily replaced. Meanwhile, both Cu and Mn showed strong PDOS contributions in CuMn₂O₄, suggesting stable Mn–O and Cu–O bonds with high electron delocalization. This structural stability and high electronic conductivity enabled Cu to efficiently and selectively replace Li in LMO, promoting the formation of Li₂SO₄ and CuMn₂O₄.

Fig. 6 (a) Total density of states for CuSO₄, Li₂SO₄, LiMn₂O₄, CuMn₂O₄, and Li₂CuO₂. The projected density of states (PDOS) of (b) CuMn₂O₄, (c) LiMn₂O₄, and (d) Li₂CuO₂.

In addition to DOS contributions, the geometric configurations of LiMn₂O₄, CuMn₂O₄, and Li₂CuO₂ were examined to assess their impact on structural stability and reaction pathway selection. Previous studies showed that the transition from metastable configurations (disordered atomic arrangement and high internal stress) to stable configurations (ordered atomic arrangement and low internal stress) reduced activation energy barriers and promoted reactions.^57,58 As shown in Fig. 6(b–d), the Cu–O and Mn–O bond lengths in CuMn₂O₄ were significantly shorter than the Li–O bonds in Li₂CuO₂, indicating a smaller unit cell for CuMn₂O₄. This suggested that CuSO₄ and LiMn₂O₄ preferentially formed the more stable CuMn₂O₄ rather than Li₂CuO₂ during sulfation roasting. Additionally, the average Mn–O bond lengths in LiMn₂O₄ and CuMn₂O₄ were extremely close, differing by only 0.018 Å. This implied a small volume expansion and low formation energy for the LiMn₂O₄-to-CuMn₂O₄ transition, making it the energetically favorable pathway. Thus, Li in LiMn₂O₄ was easily replaced by Cu during sulfation roasting, forming the more stable Li₂SO₄ and CuMn₂O₄ spinel.

Based on the above analysis, this study proposes the mechanism of selective lithium extraction from spent LMO batteries via sulfation roasting. As shown in Fig. 7, when the roasting temperature was below 500 °C, blue vitriol did not undergo significant decomposition and thus no substantial SO₂ was released. However, a small amount of lithium was still sulfated. This was attributed to the solid-phase ion exchange reaction at the contact interface between blue vitriol and LiMn₂O₄ particles. The low lithium recovery efficiency at 500 °C (Fig. 2(d)) indicated that the solid-phase ion exchange reaction was inefficient for efficient lithium recovery. This was because SO₄²⁻ exhibited poor mass transfer capacity in the solid phase, and the sulfation reaction only occurred at the contact surface without penetrating into the particle interior. When the roasting temperature exceeded 500 °C, the solid–solid exchange reaction persisted, while the gas–solid reaction gradually became the dominant process during roasting as blue vitriol progressively decomposed. Compared with SO₄²⁻, SO₂ exhibited superior diffusivity and reactivity, enabling it to penetrate deeper into LiMn₂O₄ particles. According to the geometric configuration and PDOS calculations, lithium was the most vulnerable active site in LiMn₂O₄. For this reason, SO₂ preferentially reacted with lithium to form water-soluble Li₂SO₄. Meanwhile, the structure of LiMn₂O₄ was destroyed—partial manganese was transformed into Mn₂O₃ while the remaining manganese combined with copper to form spinel-type CuMn₂O₄. Regarding the copper species, the sulfur-depleted copper either existed as CuO or participated in the formation of the CuMn₂O₄ spinel.

Fig. 7 The mechanism of selective lithium extraction from spent LMO batteries through sulfation roasting.

3.6. Recovery of lithium carbonate

Fig. S4 shows the XRD pattern of the Li₂CO₃ product. The positions and intensities of diffraction peaks matched well with the standard card of Li₂CO₃ (PDF#01-083-1454). No impurity phases were detected. The Li₂CO₃ product was further characterized using scanning electron microscopy (SEM). As indicated in Fig. S4, Li₂CO₃ exhibited irregular agglomerates composed of rod-like and plate-like crystals. Several microcracks were also observed on the surface, likely due to crystallization stress or dehydration effects during drying.⁵⁹ To further assess the purity of Li₂CO₃, the powder was digested with aqua regia and analyzed by the ICP-MS method. It was found that the purity of Li₂CO₃ reached up to 99.54 wt%, which met the battery-grade standard requirement (>99.5%, YS/T 582-2013).

3.7. Economic and environmental assessment

Fig. 8(a) shows the economic expenditures and revenues for each stage of the proposed process. The estimated profit from processing 1000 kg of LiMn₂O₄ batteries amounted to $725.06. As shown in Fig. 8(b), raw material acquisition constituted the primary cost, contributing 71.7% of the total expenditure. As shown in Fig. 8(c), the recovery of Li₂CO₃ served as the major revenue source, contributing 46.3% of the total income. Detailed economic data are provided in Tables S3–S7. It should be noted that additional government subsidies for environmentally friendly enterprises could further improve the overall profitability.

Fig. 8 (a) Economic assessment of the sulfation roasting process. (b) Proportion of each cost item. (c) Proportion of each revenue item.

This study conducted a life cycle assessment (LCA) in accordance with ISO 14040 standards to evaluate the environmental impacts of the vitriol-driven roasting for recycling LiMn₂O₄ batteries. The ISO 14040 framework divides LCA into four phases: goal and scope definition, inventory analysis, impact assessment, and interpretation. A cradle-to-gate system boundary was adopted, with the functional unit defined as the recycling of 1000 kg of LiMn₂O₄ batteries. Material inputs, energy consumption, and emissions across the recycling process were quantified using SimaPro software in conjunction with the Ecoinvent 3.9 database. Environmental impacts were assessed using the CML-IA baseline method. As summarized in Table S8, the life cycle assessment resulted in a negative carbon emission (−1.36 × 10³ kg CO₂ eq.) and energy consumption (−1.55 × 10⁴ MJ) per 1000 kg of processed LiMn₂O₄ batteries. These net negative values were largely due to the significant environmental credits achieved through closed-loop material recovery. In particular, the reuse of recycled Li₂CO₃ alone provided an emission credit of 5.63 × 10² kg CO₂ eq. and an energy saving credit of 6.17 × 10³ MJ, effectively offsetting the carbon-intensive and energy-intensive burdens associated with conventional primary lithium extraction. These results underscored the advantages of the proposed approach in reducing both carbon emissions and energy consumption.

Spent LMO batteries contained multiple hazardous substances. As the most abundant metal in the cathode, Mn inhibited plant growth if it entered the soil and harmed aquatic organisms if it entered the water system. Spent LMO batteries also contained organic solvents (e.g., electrolytes), which caused severe damage to the ecosystem through dissolution or volatilization. This study proposed a co-roasting process utilizing blue vitriol and spent LMO batteries for selective lithium recovery, simultaneously reducing raw material costs and achieving the goal of “waste-treats-waste”. Experimental results exhibited that the lithium recovery ratio exceeded 95% and the obtained Li₂CO₃ met the battery-grade standard requirement. Besides, manganese and copper remain as water-insoluble oxides, preventing their environmental migration. The process also addressed the SO₂ emission issue inherent to conventional sulfation roasting, achieving near-zero SO₂ emission during sulfation roasting. The proposed strategy realized complete recycling of spent LMO battery components, byproducts, and lithium-rich solutions. It minimized resource consumption (Reduce), maximizing material reuse (Reuse) and regeneration (Recycle). This fully aligned with the “3R” (Reduce, Reuse, Recycle) green development principles. Furthermore, the blue vitriol-driven sulfation roasting strategy was adopted for the recycling of LiCoO₂ and LiNi_0.5Co_0.2Mn_0.3O₂ batteries, achieving satisfactory lithium recovery results (Table S9). These findings demonstrated the potential applicability of the proposed strategy for recycling diverse lithium-ion battery cathodes.

4. Conclusions

In this study, an innovative blue vitriol-mediated sulfation roasting–water leaching process was developed for selective lithium extraction from spent LiMn₂O₄ batteries. This process employed blue vitriol as the sole additive for sulfation. Under optimal conditions (roasting at 700 °C, blue vitriol/LMO mass ratio of 0.7 : 1, and holding time 60 min), the Li recovery rate reached 95.18%, with a Li selectivity of 99.66%. Sample characterization revealed that LMO facilitated the decomposition of blue vitriol, generating SO₂ with strong mass transfer capability. The LMO structure was disrupted by SO₂ and converted into water-soluble Li₂SO₄ and MnSO₄. During the sulfation reaction, partial Cu and Mn formed the CuMn₂O₄ spinel at high temperature (>600 °C), which exhibited a smoother surface compared to the original spent LMO spinel structure. Density functional theory (DFT) calculations indicated that the Li–O bonds in LiMn₂O₄ were unstable, and Cu from blue vitriol had a stronger tendency to replace Li in LiMn₂O₄, validating the preferential sulfation pathway of Li at the molecular level. Under optimal conditions, SO₂ selectively sulfated all Li, while Mn and Cu remained as Mn₂O₃, CuO, and CuMn₂O₄. The resulting Li₂CO₃ product achieved a purity of 99.54%, meeting the battery-grade standards. This process demonstrated excellent economic viability and near-zero emission of harmful gases or acid/alkaline wastewater, offering a novel approach for low-value lithium-ion battery recovery. Our future research will explore scalable reactor designs and synergistic recycling of diverse LIB chemistries to further enhance circular economy practices.

Author contributions

Jiaxin Zhu—writing – original draft, investigation, formal analysis, and data curation. Yang Xu—conceptualization, writing – review & editing, project administration, and funding acquisition. Ziyao Yang—investigation and formal analysis. Qiqi Yao—investigation and formal analysis. Deyu Hua—investigation and formal analysis. Jiahao Huang—investigation and formal analysis. Xianqing Zhu—writing – review & editing and funding acquisition. Jun Li—investigation and formal analysis. Qingzhu Zhang—investigation and formal analysis.

Conflicts of interest

The authors declare no competing financial interest.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: Fig. S1, XRD pattern of the spent LMO cathode; Fig. S2, process flow diagram of selective lithium recovery from spent LMO battery; Fig. S3, modeling results of CuSO₄, Li₂SO₄, LiMn₂O₄, CuMn₂O₄, Li₂CuO₂; Fig. S4, characterization data and elemental composition of the recovered Li₂CO₃; Tables S1 and S2, characterization results of spent LMO cathode; Tables S3–S8, economic and environmental analysis using blue vitriol-driven sulfation roasting method; Table S9, the lithium recovery performance of spent NCM and LCO cathodes using blue vitriol-driven sulfation roasting method; Text S1, details of characterization. See DOI: https://doi.org/10.1039/d5gc05075h.

Acknowledgements

The authors gratefully acknowledge the financial support from the National Natural Science Foundation of China (Grant No. 52576193 and 52106168) and the Youth Innovation Program of Universities in Shandong Province (2022KJ029). The authors express their gratitude to Nannan Dong and Sen Wang of the State Key Laboratory of Microbial Technology of Shandong University for conducting the ICP-MS and SEM analysis. The authors also appreciate the testing service provided by Ceshigo Research Service (https://www.ceshigo.com) and Ukeyan (https://www.ukeyan.com).

References

1. S. Ni, J. Sheng, C. Zhang, X. Wu, C. Yang, S. Pei, R. Gao, W. Liu, L. Qiu and G. Zhou, Adv. Funct. Mater., 2022, 32, 2200682.
2. J. Yang, W. Wang, H. Yang and D. Wang, Green Chem., 2020, 22, 6489–6496.
3. L. Zhang, Y. Zhang, Z. Xu and P. Zhu, Environ. Sci. Technol., 2023, 57, 13270–13291.
4. G. Liu, Z. Liu, J. Gu, H. Yuan, Y. Wu and Y. Chen, Chem. Eng. J., 2023, 471, 144265.
5. E. Temeche, M. Yu and R. M. Laine, Green Chem., 2020, 22, 4656–4668.
6. S. Tao, H. Liu, C. Sun, H. Ji, G. Ji, Z. Han, R. Gao, J. Ma, R. Ma and Y. Chen, Nat. Commun., 2023, 14, 8032.
7. W. Ding, S. Bao, Y. Zhang, L. Ren, C. Xin, B. Chen, B. Liu, J. Xiao and X. Hou, Green Chem., 2023, 25, 6652–6665.
8. B. Qiu, M. Liu, X. Qu, F. Zhou, H. Xie, D. Wang, L. Y. S. Lee and H. Yin, Nat. Commun., 2024, 15, 6473.
9. X. Qu, B. Zhang, J. Zhao, B. Qiu, X. Chen, F. Zhou, X. Li, S. Gao, D. Wang and H. Yin, Green Chem., 2023, 25, 2992–3015.
10. X. Yang, H. Wen, Y. Lin, H. Zhang, Y. Liu, J. Fu, Q. Liu and G. Jiang, Environ. Sci. Technol., 2023, 57, 5103–5106.
11. Y. Li, W. Lv, H. Huang, W. Yan, X. Li, P. Ning, H. Cao and Z. Sun, Green Chem., 2021, 23, 6139–6171.
12. Y. Liu, X. Yang, W. Wang, W. Ma, H. Yu, C. Long, G. Li, H. Wen, Q. Liu and G. Jiang, Environ. Sci. Technol., 2025, 59, 14006–14019.
13. J. Xiao, J. Li and Z. Xu, Environ. Sci. Technol., 2019, 54, 9–25.
14. J. Lin, C. Liu, H. Cao, R. Chen, Y. Yang, L. Li and Z. Sun, Green Chem., 2019, 21, 5904–5913.
15. T. Wang, C. Xiao, Y. Qiao, S. Peng, Y. Nian and Y. Wang, Environ. Sci. Technol., 2025, 59, 16021–16033.
16. Z. Huang, R. Qiu, K. Lin, J. Ruan and Z. Xu, Environ. Sci. Technol., 2021, 55, 7643–7653.
17. H. Ali, H. A. Khan and M. Pecht, Renewable Sustainable Energy Rev., 2022, 168, 112809.
18. M. He, X. Jin, X. Zhang, X. Duan, P. Zhang, L. Teng, Q. Liu and W. Liu, Green Chem., 2023, 25, 6561–6580.
19. M. Rinne, H. Lappalainen and M. Lundström, Green Chem., 2025, 27, 2522–2537.
20. C. Pan and Y. Shen, J. Energy Chem., 2023, 85, 547–561.
21. L. Men, S. Feng, J. Zhang, X. Luo and Y. Zhou, Green Chem., 2024, 26, 1170–1193.
22. M. Wang, Q. Tan and J. Li, Environ. Sci. Technol., 2018, 52, 13136–13143.
23. Y. Yao, M. Zhu, Z. Zhao, B. Tong, Y. Fan and Z. Hua, ACS Sustainable Chem. Eng., 2018, 6, 13611–13627.
24. F. Hanna, C. Somers and A. Anctil, Environ. Sci. Technol., 2025, 59, 14432–14443.
25. B. K. Biswal, B. Zhang, P. T. M. Tran, J. Zhang and R. Balasubramanian, Chem. Soc. Rev., 2024, 53, 5552–5592.
26. L. Teng, W. Liu, X. Duan, Z. Li, C. Chen, Z. Wang, J. Yang and Q. Liu, Waste Manage., 2025, 193, 430–442.
27. L. Cheng, W. Li, Y. Li, Y. Yang, Y. Li, Y. Cheng and D. Song, J. Therm. Anal. Calorim., 2019, 135, 2697–2703.
28. D. C. Paz-Gómez, S. M. Pérez-Moreno, I. Ruiz-Oria, G. Ríos and J. P. Bolívar, Hydrometallurgy, 2021, 206, 105769.
29. M. He, W. Cao, L. Teng, W. Liu, S. Ji, W. Yu, C. Ding, H. Wu and Q. Liu, J. Colloid Interface Sci., 2024, 663, 930–946.
30. M. Tang, A. Yuan, H. Zhao and J. Xu, J. Power Sources., 2013, 235, 5–13.
31. R. Yaokawa, H. Sugimoto, S. Kosaka, T. Nonaka and R. Kitagawa-Yogo, Environ. Sci. Technol., 2025, 59, 9600–9609.
32. L. Yang, H. Zhang, F. Luo, Y. Huang, T. Liu, X. Tao, G. Yang, X. Luo and P. Shao, Resour., Conserv. Recycl., 2024, 203, 107460.
33. Y. Yang, X. Meng, H. Cao, X. Lin, C. Liu, Y. Sun, Y. Zhang and Z. Sun, Green Chem., 2018, 20, 3121–3133.
34. Z. Gao, Y. Liu, Z. M. El Bahy, G. Chen, M. H. Helal, B. Lu, J. Han and J. Zhou, Adv. Funct. Mater., 2025, 35, 2503674.
35. L. Teng, W. Liu, M. He, Z. Wang and Q. Liu, Journal of Environmental Chemical Engineering., 2024, 12, 113521.
36. Y. Wang, Q. Zhang, X. Bai, P. Mu, X. Zhang, W. Wang, Z. Tan, Z. Zhao and D. Pan, Chem. Eng. J., 2023, 467, 143441.
37. X. Yongyao, K. Naoki and Y. Masaki, J. Power Sources, 2000, 90, 135–138.
38. A. M. Gadalla, Int. J. Chem. Kinet., 1984, 16, 655–668.
39. L. Mao, T. Ali, C. Huang and N. Z. Muradov, Int. J. Hydrogen Energy, 2011, 36, 5822–5827.
40. W. Fu, L. Meng and J. Qu, Metals, 2024, 14, 1075.
41. X. Cai, Q. Liu, C. Xia, D. Shan, J. Du and J. Chen, Environ. Sci. Technol., 2015, 49, 9264–9272.
42. L. Hughes, A. Roy, C. Downing, M. P. Browne, A. Zhussupbekova, I. V. Shvets and V. Nicolosi, Adv. Funct. Mater., 2023, 33, 2214883.
43. J. Zhang, B. Shen, Z. Hu, M. Zhen, S. Guo and F. Dong, Appl. Catal., B, 2021, 296, 120376.
44. M. Khan, E. Suljoti, A. Singh, S. A. Bonke, T. Brandenburg, K. Atak, R. Golnak, L. Spiccia and E. F. Aziz, J. Mater. Chem. A, 2014, 2, 18199–18203.
45. T. Okazaki, T. Yoshioka, Y. Kusakabe, T. Yamamoto and A. Kitada, Solid State Commun., 2011, 151, 1749–1752.
46. D. R. Baer, G. E. McGuire, K. Artyushkova, C. D. Easton, M. H. Engelhard and A. G. Shard, J. Vac. Sci. Technol., A, 2021, 39, 021601.
47. G. Greczynski and L. Hultman, Prog. Mater. Sci., 2020, 107, 100591.
48. Z. Liang, C. Wang, G. Li, H. Zhuo, L. Guo, H. Liu, Z. Jin, Z. Ren and J. Wang, ACS Appl. Energy Mater., 2025, 8, 4166–4175.
49. D. Yang, F. Dong, J. Wang, Z. Tang and J. Zhang, Environ. Sci.:Nano, 2024, 11, 1926–1947.
50. Z. Wu, R. Jin, H. Wang and Y. Liu, Catal. Commun., 2009, 10, 935–939.
51. M. Baazizi, M. Karbak, M. Aqil, S. Sayah, M. Dahbi and F. Ghamouss, ACS Appl. Mater. Interfaces, 2023, 15, 40385–40396.
52. R. E. Warburton, F. C. Castro, S. Deshpande, K. E. Madsen, K. L. Bassett, R. Dos Reis, A. A. Gewirth, V. P. Dravid and J. Greeley, ACS Appl. Mater. Interfaces, 2020, 12, 49182–49191.
53. B. Saravanakumar, S. M. Lakshmi, G. Ravi, V. Ganesh, A. Sakunthala and R. Yuvakkumar, J. Alloys Compd., 2017, 723, 115–122.
54. A. A. Anber, A. S. Hameed, A. M. Rheima and J. M. Dhabab, Case Stud. Chem. Environ. Eng., 2024, 10, 101003.
55. Y. Wang, S. Wang, Y. Guo, D. Xu, Y. Gong, X. Tang and H. Ma, Ind. Eng. Chem. Res., 2012, 51, 16573–16579.
56. M. Ullah, E. Ahmed, F. Hussain, A. M. Rana and R. Raza, Appl. Surf. Sci., 2015, 334, 40–44.
57. R. A. Van Santen, M. Neurock and S. G. Shetty, Chem. Rev., 2009, 110, 2005–2048.
58. K. K. Mentel, A. Serra, P. E. Abreu and L. G. Arnaut, Nat. Commun., 2018, 9, 2903.
59. N. Chen, E. Zhou, D. Duan and X. Yang, Korean J. Chem. Eng., 2017, 34, 2748–2755.

---