Scalable Hydrometallurgical Process for LFP Battery Recycling with Selective Lithium Extraction and Iron Stabilization

Abstract

The expanding electric vehicle market is driving a critical demand for lithium-ion battery materials, positioning spent batteries as a strategic urban resource. Conventional hydrometallurgical studies are typically constrained by low solid-to-liquid (S/L) ratios (e.g., 1:50 g/mL), which limits their industrial relevance. This work adapted and optimized the MONOLITHOS-patented HCl–H₂O₂ leaching system to achieve in situ oxidative delithiation and selectively extract lithium (Li) from spent lithium iron phosphate (LFP) black mass under intensified S/L ratios (1:10 to 1:2 g/mL) relevant to scalable applications. Leaching parameters were optimized using response surface methodology based on a Box–Behnken design. High solids loadings introduced mass-transfer limitations and proton depletion, while LFP delithiation was governed by competitive iron dissolution. Under optimized ambient leaching conditions at intensified pulp density, selective leaching achieved 86.0% Li extraction efficiency while suppressing iron dissolution to 19.2% through solid-state FePO₄ fixation. Downstream purification at a controlled acidic pH precipitated 99.7% of the dissolved iron, with minimal lithium co-precipitation (<0.55%). Subsequent oxidative polishing removed residual iron, and thermal preconcentration enabled carbonate precipitation, yielding technical-grade Li₂CO₃ with an estimated mass-balance purity of 97.3%. The results demonstrate an efficient, low-energy, intensified route for sustainable LFP recycling at industrially relevant pulp densities.

1. Introduction

Lithium-ion batteries (LIBs) are widely deployed in electric vehicles, portable electronics, and large-scale energy storage systems due to their superior energy density, extended cycle life, and environmental compatibility within the global transition toward renewable energy [1,2]. Global LIB shipments reached 1545.1 GWh in 2024 and are projected to exceed 5000 GWh by 2030 [3]. Among prevailing LIB cathode chemistries (including LiNi_xMn_yCo_1−x−yO₂ [NMC], LiMn₂O₄ [LMO], LiCoO₂ [LCO], and LiFePO₄ [LFP]), LFP is increasingly prioritized for electric mobility owing to its exceptional thermal stability, structural reversibility, and lower raw material costs [4,5]. With global electric vehicle deployment projected to reach 145 million units by 2030, demand for LFP cathodes is accelerating rapidly [6]. However, LFP batteries typically reach their end-of-life (EoL) within 5–10 years [7], generating vast volumes of spent battery waste that pose severe resource management, human health, and environmental challenges if unmanaged [8]. EoL battery management devoid of good recycling practices leads to hazardous electrolytes (LiPF₆) decomposition and phosphorus leaching, posing great ecological and public health risks [9,10,11]. Nevertheless, spent LFP cathode black mass represents a high-grade secondary resource rich in lithium (2–5 wt%) alongside concentrated iron and phosphate (P₂O₅ > 40 wt%) [12,13,14]. Efficient recycling of spent LFP batteries is therefore imperative to mitigate ecological hazards, secure critical mineral supply chains, and advance circular economy principles. Unlike cobalt- and nickel-rich cathode chemistries (e.g., NMC), where recovery is driven by high-value transition metals, the primary economic value in spent LFP black mass is concentrated almost in lithium, while iron and phosphorus constitute low-value matrix components (Table 1). Consequently, LFP recycling requires a distinct hydrometallurgical strategy focused on selective lithium delithiation while keeping iron in the solid phase [15].

Table 1. Elemental composition (wt%) and market value benchmarks of spent LFP vs. NMC cathode materials (data adapted from Gerold et al. [16]).

Element	LFP (wt%)	NMC (wt%)	Market Value Benchmark
Lithium (Li)	2.78	4.76	High; Critical Raw Material
Iron (Fe)	21.53	–	Low value (primary fixation target)
Phosphorus (P)	11.81	–	Strategic (agricultural/chemical use)
Nickel (Ni)	–	7.94	High price volatility
Cobalt (Co)	–	3.97	High cost/elevated supply risk
Manganese (Mn)	–	29.37	Major structural component in NMC; low market value

Currently, spent LFP recycling is predominantly categorized into pyrometallurgical processing, direct cathode regeneration, and hydrometallurgical extraction [17]. Pyrometallurgical smelting operates via simple workflows and minimizes aqueous waste generation; however, it requires extreme thermal energy input (>1200 °C) to break down the LFP olivine crystal matrix. Furthermore, pyrometallurgy locks valuable lithium into the slag phase, requiring capital-intensive secondary recovery steps while releasing hazardous off-gases that necessitate complex scrubbing systems [18,19]. Direct regeneration seeks to repair degraded cathodes by restoring local stoichiometry (Li, Fe, and P) in situ via solid-state or hydrothermal re-lithiation [20,21]. Despite its low chemical consumption, direct regeneration is highly sensitive to initial cathode degradation history, surface impurity contamination, and structural lattice defects, rendering commercial-scale implementation difficult for heterogeneous waste feeds [22]. Conversely, hydrometallurgy offers superior extraction efficiencies, exceptional product purity, and significantly lower energy requirements [23]. This route relies on aqueous chemical leaching to solubilize cathode constituents from black mass, followed by liquid-phase separation, purification, and salt crystallization [24].

To optimize hydrometallurgical delithiation, various lixiviants have been investigated. Organic acids such as acetic acid (CH₃COOH) [25], formic acid (CH₂O₂) [26], and p-toluenesulfonic acid (TSA) [27] demonstrate selective leaching capabilities; however, their industrial deployment is bottlenecked by sluggish kinetics, high reagent costs, and unwanted ferric-organic chelation that lowers Li/Fe separation factors [28,29]. Alternatively, strong chemical oxidants including sodium persulfate (Na₂S₂O₈) [30], sodium hypochlorite (NaClO) [31], and ammonium persulfate ((NH₄)₂S₂O₈) [32] can selectively extract Li⁺ ions without collapsing the core crystal lattice. Nevertheless, these oxidants suffer from poor thermal stability, rapid self-decomposition, and high operational costs [33]. Consequently, inorganic mineral acids such as sulfuric acid (H₂SO₄) [34], phosphoric acid (H₃PO₄) [35], and hydrochloric acid (HCl) [36] remain the industry standard due to their high reactivity, low cost, and feedstock adaptability. For instance, Cong et al. [37] reported 95.23% Li recovery from LFP black mass using a 0.315 M H₂SO₄-H₂O₂ mixture; however, achieving high yield required an elevated operating temperature (62 °C). Similarly, Chen et al. [38] achieved 100% Li extraction with only 2.7% Fe co-dissolution using H₃PO₄, but the system relied on high chemical consumption. A major limitation of conventional inorganic leaching frameworks is their reliance on high thermal inputs (60–90 °C) to overcome kinetic barriers, introducing a substantial energy penalty.

To overcome these kinetic and energetic barriers, the combination of hydrochloric acid (HCl) and hydrogen peroxide (H₂O₂), functioning via synergistic proton-exchange and oxidative chemistry, has been shown to be highly effective among leaching systems [38]. The chloride medium promotes the formation of distinct chloro-complexes when compared with the sulfate medium [39]. In a comparative thermodynamic study, Mohammad Zadeh et al. [40] reported superior selectivity for HCl (Li/Fe > 100) relative to H₂SO₄ at 38 °C. Within the HCl–H₂O₂ system, controlled oxidation–reduction potential (ORP) enables in situ oxidation of Fe²⁺ to Fe³⁺ inside the olivine lattice. This drives selective release of Li⁺ into solution while retaining iron as an insoluble solid-state ferric phosphate (FePO₄) phase. The mechanism permits selective lithium extraction at ambient temperature (25 ± 2 °C), thereby eliminating the need for external heating.

Despite these advantages, process intensification and industrial scalability remain primary bottlenecks in hydrometallurgical recycling. As summarized in Table 2, recent selective LFP delithiation studies (2022–2026) have demonstrated high lithium extraction efficiencies using a range of mineral and organic acid-based lixiviants. Nevertheless, many of these processes remain limited to relatively dilute pulp densities (50–186.9 g/L) or rely on elevated temperatures (60–250 °C) to overcome kinetic barriers. Attempting to scale up by simply increasing batch pulp density (>200 g/L) often induces a sharp drop in metal yield due to boundary-layer mass transfer limitations and proton starvation. Furthermore, non-linear interactions between acid strength, oxidant dosage, and solid loading render conventional one-factor-at-a-time experimental approaches inefficient. Response surface methodology (RSM) based on a Box–Behnken design (BBD) provides a rigorous statistical framework to model non-linear parameter interactions, evaluate response surface curvature, and identify process optima [41].

The main objective of this study is to adapt and optimize the Monolithos versatile HCl-H₂O₂ leaching technology [42] for the selective delithiation of non-thermally treated spent LFP black mass under process-intensified conditions. This work addresses this literature gap by systematically evaluating selective leaching mechanisms across high S/L ratios (10–50% w/v, corresponding to 100–500 g/L). RSM is used to map parameter interactions (HCl concentration, H₂O₂ peroxide dosage, and S/L ratio) and establish operational boundaries for lithium extraction and iron suppression. Finally, an integrated downstream purification sequence combining NaOH-H₂O₂ precipitation, oxidative solution polishing, thermal preconcentration, and crystallization is validated to produce high-purity lithium carbonate (Li₂CO₃) precursors for closed-loop battery manufacturing.

Table 2. Selected recent lixiviant systems, leaching parameters, and lithium efficiencies for the selective leaching of lithium from spent LFP black mass [16,43,44,45,46,47,48,49,50,51].

Reference	Acid System	Acid Conc.	Oxidant	Oxidant Dosage	Temp. (°C)	Time (Min)	S/L Ratio (g/L)	Li Extraction (%)
Gerold et al. (2022) [16]	H2SO4	2.0 M	H2O2	0.02 vol%	30	240	96.88	3.2–4.2 g/L *
Ren et al. (2024) [44]	HCOOH	0.5 M	H2O2	2.5 vol%	20	30	100	>97.00%
Moschovi et al. (2023) [45]	HCl	0.60 M	NaClO	ClO−/Li = 0.6 mol	15	20	71.4	>95.00%
Chen et al. (2015) [46]	H2SO4	0.4 M	H2O2	H2O2/Li = 2.5 mol	60	120	~100.0	97.95%
Ou et al. (2024) [47]	HCOOH	0.8 M	H2O2	8.0 vol%	60	60	50	99.90%
Zhang et al. (2024) [48]	(NH4)2SO4	1.0 M	H2O2	4.0 vol%	250	10	80	97.70%
Xu et al. (2025) [49]	HNO3	HNO3/Li = 2.0 mol	—	—	120	120	100	99.50%
Yang et al. (2019) [43]	H2SO4	0.315 M	H2O2	H2O2/Li = 2.13 mol	62	90	96.87	95.23%
Liu et al. (2025) [50]	H2SO4	0.9 M	H2O2	H2O2/Li = 2.1 mol	45	10	186.9	98.72%
Zou et al. (2024) [51]	HCl/FeCl3	5.0 M	FeCl3	Fe3+/Li+ = 1.2 mol	35	30	166.7	97.00%
This Work	HCl–H2O2	2.34 M	H2O2	3.0 vol%	25	60	32.86% w/v (328.6 g/L)	86.00%

* Reported as concentration yield in pregnant leach liquor rather than percentage extraction.

Table 2. Selected recent lixiviant systems, leaching parameters, and lithium efficiencies for the selective leaching of lithium from spent LFP black mass [16,43,44,45,46,47,48,49,50,51].

Reference	Acid System	Acid Conc.	Oxidant	Oxidant Dosage	Temp. (°C)	Time (Min)	S/L Ratio (g/L)	Li Extraction (%)
Gerold et al. (2022) [16]	H2SO4	2.0 M	H2O2	0.02 vol%	30	240	96.88	3.2–4.2 g/L *
Ren et al. (2024) [44]	HCOOH	0.5 M	H2O2	2.5 vol%	20	30	100	>97.00%
Moschovi et al. (2023) [45]	HCl	0.60 M	NaClO	ClO−/Li = 0.6 mol	15	20	71.4	>95.00%
Chen et al. (2015) [46]	H2SO4	0.4 M	H2O2	H2O2/Li = 2.5 mol	60	120	~100.0	97.95%
Ou et al. (2024) [47]	HCOOH	0.8 M	H2O2	8.0 vol%	60	60	50	99.90%
Zhang et al. (2024) [48]	(NH4)2SO4	1.0 M	H2O2	4.0 vol%	250	10	80	97.70%
Xu et al. (2025) [49]	HNO3	HNO3/Li = 2.0 mol	—	—	120	120	100	99.50%
Yang et al. (2019) [43]	H2SO4	0.315 M	H2O2	H2O2/Li = 2.13 mol	62	90	96.87	95.23%
Liu et al. (2025) [50]	H2SO4	0.9 M	H2O2	H2O2/Li = 2.1 mol	45	10	186.9	98.72%
Zou et al. (2024) [51]	HCl/FeCl3	5.0 M	FeCl3	Fe3+/Li+ = 1.2 mol	35	30	166.7	97.00%
This Work	HCl–H2O2	2.34 M	H2O2	3.0 vol%	25	60	32.86% w/v (328.6 g/L)	86.00%

* Reported as concentration yield in pregnant leach liquor rather than percentage extraction.

2. Materials and Methods

2.1. Chemicals and Reagents

The leaching reagents used in this research, hydrochloric acid (HCl, 37% w/w aqueous solution, ACS grade, PanReac AppliChem, Castellar del Vallès, Spain) and hydrogen peroxide (H₂O₂, 35% w/w aqueous solution, Solvay S.A., Brussels, Belgium), were of analytical grade. HCl was used to assess the leaching performance of the LFP cathode material, while H₂O₂ served as an oxidizing agent. Sodium hydroxide (NaOH, analytical grade) was employed for pH adjustment during Fe precipitation, while anhydrous sodium carbonate (Na₂CO₃, analytical grade) served as the precipitant for Li₂CO₃ synthesis. All chemicals were used as received without further purification. Finally, all solutions were prepared and diluted using deionized water.

2.2. Feed Characterization

The primary feedstock used in this study was non-thermally treated black mass derived from a spent LiFePO₄ (LFP) battery, supplied by Monolithos Catalysts and Recycling Ltd. (Athens, Greece). To ensure physical and chemical consistency across subsequent leaching experiments, the raw black mass was mechanically sieved to achieve a uniform particle size fraction of <150 μm (with characteristic distribution metrics $d_{50}\approx 45 \mathsf{\mu }\mathrm{m}$ and $d_{90}\approx 110 \mathsf{\mu }\mathrm{m}$). This fine granulometry ensures high interfacial surface area, minimizing unreacted core diffusion barriers across the solid particles. To determine the head grade of the LFP black mass, total acid dissolution was executed on representative samples. Complete dissolution was achieved using an initial sample mass of 0.50 g, digested in aqua regia at 95 °C for 2 h. Following digestion, the solution was quantitatively diluted to a final volume of 100 mL using deionized water. The resulting leachate was analyzed with Inductively Coupled Plasma Optical Emission Spectroscopy (ICP-OES, Thermo Fisher Scientific, Waltham, MA, USA, iCAP 6500 Series) in radial view mode. Specific atomic emission lines were selected to minimize spectral overlaps, comprising Li (610.3 nm) and Fe (218.7 nm). The elemental concentrations of the LFP black masses are summarized in Table 3.

Table 3. Primary metal components in the cathode materials of the LFP LIBs.

Element	Concentration (mg/kg)	Weight Fraction (wt%)
Fe	321,236	32.12
Li	48,069	4.81

2.3. Leaching Experiments

All batch leaching experiments were performed in a 250 mL glass beaker equipped with a cover to minimize evaporative losses and maintain precise control over reaction conditions. The selective leaching of lithium from the LFP black mass was performed at room temperature (25 ± 2 °C) to evaluate low-energy recovery. A measured volume of HCl (2.0, 2.5, and 3.0 M) was added to the beaker, followed by LFP powder at solid-to-liquid (S/L) pulp densities of 10%, 30%, and 50% w/v (corresponding to 100, 300, and 500 g/L, or solid-to-liquid mass-to-volume ratios of 1:10, 1:3.33, and 1:2 g/mL, respectively). The mixture remained under stirring to maintain a uniform suspension and prevent mass-transfer limitations. The leaching process was formally initiated by the addition of H₂O₂ (1.0–3.0% v/v), with t = 0 defined upon completing the oxidant addition. Here, H₂O₂ acts as an oxidant, facilitating delithiation from the olivine structure by oxidizing Fe²⁺ to Fe³⁺, thereby driving the selective dissolution of Li while retaining the iron-phosphate matrix in the solid phase.

After maintaining the reaction for 1 h, the slurry was separated by vacuum filtration using 3–5 μm qualitative filter paper (Shaanxi Syvet, Xi’an, China). The resulting filter cake was washed with 1.0 M HCl to recover entrained lithium ions and dried at 90 °C overnight. The pregnant leach solution (PLS) was analyzed for Li and Fe via ICP-OES (calibrated with 99.9% multi-element standards) to determine the selectivity index, while the solid residue was reserved for further characterization. Crystalline phase identification of the synthesized lithium carbonate precipitate was carried out by X-ray diffraction (XRD) using Cu–Kα radiation (λ = 1.54060 Å) in a coupled 2θ/θ scanning geometry. To further assess the degree of selective leaching of Li in the leaching system, the leaching selectivity was introduced, and it is given by Equation (1):

$${\mathit{L}\mathit{e}\mathit{a}\mathit{c}\mathit{h}\mathit{i}\mathit{n}\mathit{g} \mathit{s}\mathit{e}\mathit{l}\mathit{e}\mathit{c}\mathit{t}\mathit{i}\mathit{v}\mathit{i}\mathit{t}\mathit{y} (\mathit{S}}_{\mathit{L}\mathit{i}})={\displaystyle \frac{{\mathit{\rho }}_{\mathit{L}\mathit{i}}}{{\mathit{\rho }}_{\mathit{L}\mathit{i}}+{\mathit{\rho }}_{\mathit{F}\mathit{e}}}}\times \mathbf{100}\mathbf{\%}$$

(1)

where ρ_i (mol/L) denotes the molar concentration of lithium and iron metal ions in the leaching solution.

The leaching dissolution efficiency (η%) of each element (Li or Fe) was calculated using Equation (2):

$$\mathit{\eta }\mathit{\%}={\displaystyle \frac{{\mathit{C}}_{\left(\mathit{l}\right)}\times {\mathit{V}}_{\left(\mathit{f}\right)}}{{\mathit{C}}_{\left(\mathit{s}\right)}\times {\mathit{M}}_{\left(\mathit{i}\right)}}}\times \mathbf{100}$$

(2)

where C_(l) is the element concentration in the leach liquor, V_(f) is the volume of the final leach liquor, C_(s) is the element concentration in the black mass, and M_(i) is the initial mass of black mass input fed into the beaker.

2.4. Design of Experiments (DoE)

To systematically establish predictive mathematical models for the selective extraction of Li from EoL LFP cathode material, an RSM based on a three-factor BBD was implemented. This statistical framework was utilized to model the non-linear relationship between key process parameters and metal recoveries while minimizing the experimental footprint. The independent variables investigated were HCl concentration (Factor A), H₂O₂ dosage (Factor B), and solid-to-liquid ratio (Factor C). These factors were evaluated at three coded levels: low (−1), center (0), and high (+1). The parameter level codes and the experimental ranges derived for the BBD leaching optimizations are detailed in Table 4.

Table 4. Parameter level codes and values used in LFP leaching BBD.

Factor	Low (−1)	Mid (0)	High (+1)
A: HCl Conc [M]	2	2.5	3
B: H2O2 Dosage [vol%]	1	2	3
C: S/L Ratio [w/v%] (g/L)	10% (100 g/L)	30% (300 g/L)	50% (500 g/L)

The Minitab software (Minitab^® LLC, State College, PE, USA, version 22.4.0) was used to design the three-factor and three-level experiments. With the statistical analysis provided by the software, the regression and graphical analysis were obtained. To quantify the leaching recovery (Y), a second-order polynomial model was proposed to describe the system’s behavior, accounting for linear effects, mutual interactions, and quadratic curvature represented as Equation (3):

$$\mathbf{Y}={\mathsf{\beta }}_{\mathbf{o}}+{\displaystyle \sum _{\mathbf{i}=1}^{\mathbf{K}}}{\mathsf{\beta }}_{\mathbf{i}}{\mathbf{x}}_{\mathbf{i}}+{\displaystyle \sum _{\mathbf{i}=1}^{\mathbf{K}}}{\mathsf{\beta }}_{\mathbf{i}\mathbf{i}}{\mathbf{x}}_{\mathbf{i}}^2+{\displaystyle \sum _{\mathbf{i}=1}^{\mathbf{K}-1}}{\displaystyle \sum _{\mathbf{j}=2}^{\mathbf{K}}}{\mathsf{\beta }}_{\mathbf{i}\mathbf{j}}{\mathbf{x}}_{\mathbf{i}}{\mathbf{x}}_{\mathbf{j}}+\mathsf{\epsilon }$$

(3)

where Y is a response variable, k is the number of independent variables, ε is the unknown error constant, β₀ is a constant, and β_i is the regression coefficient. The statistical analysis of models and the quantification of variable significance have been conducted by ANOVA (analysis of variance).

To assess experimental reproducibility and calculate pure experimental error, five independent replicate runs were conducted at the design center point (2.5 M HCl, 2 vol% H₂O₂, and 30% w/v solid loading). Furthermore, optimal condition validation trials were executed in triplicate ($\mathit{n}=3$), and values are reported as mean ± standard deviation.

2.5. Purification of LFP Leachate

Following the leaching of LFP black mass using a HCl-H₂O₂ system, the resulting leachate was processed in a subsequent purification stage to remove iron by sequential pH adjustment with NaOH. Precipitation experiments were conducted in 250 mL beakers stirred with a PTFE-coated magnetic stirrer bar (Fisherbrand, Pittsburgh, PA, USA, 30 × 60, octagonal) at a constant rate of 300 rpm. For each experimental run, 100 mL of LFP leachate was introduced into the beaker, which was equipped with a glass cover to minimize evaporation losses. A known volume above the stoichiometric requirement of 1% w/v H₂O₂ was initially added to the PLS for pretreatment and to ensure sufficient oxidation of Fe²⁺ to Fe³⁺.

The PLS was preheated to 50 °C before the precipitating agent (3.0 M NaOH) was added dropwise. A thermometer was utilized to monitor the solution, maintaining a steady reaction temperature of 60 °C. The experiments investigated the effect of pH at target values of 2.5, 3, 3.5, and 4. The precipitation start time for all runs was defined as the moment the target pH was reached. Once attained, the reaction was maintained for a duration of 30 min. Upon completion, the slurry was subjected to vacuum filtration using a hydrophilic PTFE membrane filter (0.22 μm pore size) to separate the residue from the purified PLS. To maximize lithium recovery and ensure the accuracy of the mass balance, the obtained residue was thoroughly washed with deionized water to displace any entrained lithium-bearing mother liquor. The filter cakes were further dried in an oven at 80 °C for a minimum of 24 h until a constant weight was achieved. The metal concentrations in the filtrate were determined via ICP-OES (Thermo Fisher Scientific, Waltham, MA, USA), and the precipitate was subjected to characterization. The precipitation percentage (P*) was calculated using Equation (4) [52]:

$${\mathit{P}}^{\mathbf{\ast }}={\displaystyle \frac{{\mathit{C}}_{\mathit{o}}{\mathit{V}}_{\mathit{o}}-\mathit{C}\mathit{V}}{{\mathit{C}}_{\mathit{o}}{\mathit{V}}_{\mathit{o}}}}\times \mathbf{100}\mathit{\%}$$

(4)

where C₀ is the element concentration in the solution before precipitation, C is the solution concentration after precipitation, V₀ is the solution volume before precipitation, and V is the volume of the solution after precipitation. This equation expresses P* as the percentage of elements precipitated during each step.

2.6. Li₂CO₃ Synthesis

Following optimized leaching and impurity removal, lithium was recovered from the purified pregnant leach solution (PLS) as lithium carbonate (Li₂CO₃). The purified leachate was first preconcentrated by continuous heating and then treated with a stoichiometric excess of saturated Na₂CO₃ solution at 90 °C under continuous stirring (300 rpm) for 1 h to induce precipitation. After a 30 min settling period, the Li₂CO₃ crystals were collected by vacuum filtration using qualitative filter paper (0.45 µm pore size, Shaanxi Syvet). The recovered solids were thoroughly washed with hot deionized water to remove residual sodium ions and subsequently oven-dried to constant mass at 85 °C for 1 h prior to characterization.

3. Results and Discussion

3.1. Statistical Modeling and ANOVA Analysis

Using a BBD experiment conducted at room temperature, the leaching was fixed at 60 min within the selective leaching range. The leaching experimental results for lithium (Li) and iron (Fe) were analyzed using RSM. A total of 17 experimental runs were conducted to facilitate the establishment of a statistical relationship between the independent leaching parameters HCl concentration (A), H₂O₂ dosage (B), and S/L ratio (C), and the response (Y) in percentage dissolution, represented by second-order polynomial quadratic equations (Table 5). The empirical model for Li dissolution and Fe co-dissolution is presented in Equations (5) and (6):

$$Y_{Li}=126.0+3.0A-27.3B-0.923C-5.90A^2+2.77B^2-0.03869C^2+4.0AB+0.775AC+0.275BC$$

(5)

$$Y_{Fe}=506-200.3A-108.9B-5.74C+30.0A^2+10.75B^2+0.04687C^2+20.50AB+0.0525AC+0.025BC$$

(6)

where Y_Li and Y_Fe represent the predicted percentage extraction of lithium and co-dissolution of iron, respectively; A denotes the hydrochloric acid concentration (M); B represents the hydrogen peroxide dosage (vol%); and C represents the solid-to-liquid pulp density (% w/v).

Table 5. The Box–Behnken design of the variables and response for Li and Fe.

Run Order	Standard Order	HCl Conc. [M]	H2O2 Dosage [vol%]	S/L Ratio [w/v%]	Li Diss. (%)	Fe Co-Diss. (%)
1	4	3	3	30	92	44
2	6	3	2	10	78	85
3	11	2.5	1	50	49	36
4	3	2	3	30	86	13
5	13	2.5	2	30	90	33
6	8	3	2	50	66	32
7	10	2.5	3	10	85	84
8	9	2.5	1	10	93	99
9	16	2.5	2	30	83	29
10	5	2	2	10	86	93
11	7	2	2	50	43	19
12	12	2.5	3	50	63	23
13	17	2.5	2	30	83	30
14	2	3	1	30	83	65
15	15	2.5	2	30	82	31
16	1	2	1	30	85	75
17	14	2.5	2	30	88	32

The model equation illustrates the impact of all linear terms (A, B, and C), quadratic terms (A², B², and C²), and two-way interactions (AB, AC, and BC) between the process conditions and responses. To ensure statistical integrity, the reliability of the regression models was examined via Analysis of Variance (ANOVA), as detailed in Table 6. Significance testing was performed at a 95% (α = 0.05) confidence level; models for both Li (F = 31.18, p < 0.00) and Fe (F = 22.98, p = 0.01) were found to be highly statistically significant.

Table 6. Analysis of variance (ANOVA) for Li and Fe dissolution efficiency (%).

Source		Li Dissolution		Fe Co-Dissolution
	DOF	F-Value	p-Value	F-Value	p-Value
Model	9	31.18	0	22.98	0
A: HCl (M)	1	3.8	0.09	1.40	0.28
B: H2O2 (vol%)	1	2.7	0.15	25.59	0.00
C: S/L Ratio (%)	1	154.25	0	130.86	0
AB Interaction	1	1.35	0.28	6.98	0.03
AC Interaction	1	20.25	0.00	1.83	0.22
BC Interaction	1	10.2	0.015	0.02	0.90
A2	1	0.77	0.41	3.94	0.09
B2	1	2.73	0.14	8.09	0.03
C2	1	84.99	0	24.6	0.00
Residual Error	7	-	-	-	-
Lack-of-Fit	3	0.85	0.54	54.83	0.00
Pure Error	4	-	-	-	-
Total	16	-	-	-	-

Note: Degrees of freedom (DOF); F-value denotes the ratio of mean squares of the model term to residual mean square error; p-value represents the probability of observing the test results under the null hypothesis (statistical significance established at p < 0.05); Residual error represents the total unexplained variance remaining after model fitting, partitioned into lack-of-fit and pure experimental error.

For Li extraction, the R² value of 97.57% and a Predicted R² of 82.55% indicate high model adequacy and strong predictive capability, and can effectively explain the process being studied (Table 7). Conversely, the Fe model exhibited a significant lack-of-fit (p = 0.001). While a significant lack-of-fit generally indicates that a regression equation possesses mathematical limitations in modeling data variance, this statistical outcome is physically justified by the underlying chemistry of the system.

Table 7. Adequacy of mathematical models established to predict metal solubilization from LFP black mass.

Model Fitness	Li	Fe
R2	97.57%	96.73%
R2 (adj)	94.44%	92.52%
R2 (pred)	82.55%	48.74%

The disparity observed between the high coefficient of determination (R² = 96.73%) and the lower predicted coefficient (R²_pred = 48.74%), accompanied by a significant lack-of-fit (p = 0.001), indicates that the empirical second-order polynomial model exhibits reduced predictive extrapolation capabilities for iron co-dissolution across boundary conditions. From a physical chemistry perspective, this mathematical discrepancy does not stem from experimental error, but reflects the severe, non-linear phase-transition boundaries inherent to the system. Iron behavior is governed by simultaneous, competitive reaction pathways: acid-driven matrix dissolution versus peroxide-induced oxidative precipitation of solid-state FePO₄. At the critical redox/pH boundary (Figure 6), iron shifts abruptly between complete solubility and insoluble ferric phosphate fixation, introducing sharp gradient transitions that second-order polynomial regression models cannot smoothly accommodate. Nevertheless, the high adjusted coefficient of determination (R²_adj = 92.52%) confirms that over 92% of the internal response variance is adequately explained, allowing the response surface to reliably identify the multi-variable optimization optimum within the evaluated experimental domain.

Nevertheless, the high adjusted coefficient of determination ($R_{\mathrm{adj}}^2=92.52\mathrm{\%}$) confirms that over 92% of the response variance is adequately explained by the fitted RSM. Thus, the model retains sufficient predictive capacity to reliably establish optimal conditions and define the boundaries of the selective leaching design space.

According to the deviation between the Li dissolution predicted through the model and the actual experimental results, as shown in Figure 1a, the data points exhibit excellent linearity along the 45° diagonal, confirming that the 17 sets of experimental runs are normally distributed and free from major systematic errors. Again, Figure 1b displays the relationship between the number of experiments and the internally studentized residuals.

Figure 1. (a) Linear plots of model-predicted vs. actual experimental dissolution efficiencies. (b) The plot of internally studentized residuals.

The residuals are distributed over the interval (−3, +5), indicating a symmetric distribution about zero and confirming that the regression model has been correctly established and satisfies the statistical assumption of homoscedasticity. The Pareto charts for the ANOVA F-values, as shown in Figure 2, reveal the level of importance of the various individual parameters and their interactions with Li and Fe, respectively. For Li extraction, the hierarchy of significance follows the order C > CC > AC > BC, showing that lithium release is controlled primarily by the solid–liquid ratio rather than linear acid strength or oxidant concentration. While H₂O₂ dosage is significant for Fe, standalone HCl concentration is statistically non-significant for both elements across the evaluated range. In the LFP matrix, Li⁺ is liberated once Fe is oxidized or the olivine structure is destabilized, rendering delithiation non-acid-limited within the tested domain. The strong significance of C and CC indicates that Li extraction increases sharply at lower pulp densities but exhibits diminishing returns at very low S/L ratios. Furthermore, the significant AC and BC interaction terms demonstrate that the beneficial effects of acid and oxidant become observable only when the S/L ratio is low enough to permit adequate interfacial mass transfer; at elevated solid loadings, severe mass-transfer resistance and localized reagent starvation dominate the kinetics. In contrast, for Fe dissolution, the hierarchy of statistical influence is given by C > B > CC > B² > AB, confirming that oxidant dosage and its synergistic interaction with acid strength are critical in governing redox potential and driving solid-state FePO₄ fixation.

Figure 2. The Pareto chart of the standardized effects for Li dissolution and Fe co-dissolution.

The individual plots showing the individual influences of the input parameters for Li dissolution and Fe co-dissolution were evaluated using main effects plots, as shown in Figure 3, respectively. These one-factor plots assume all other factors are constant while showing the overall effect of a certain process condition on the metal dissolution. The results in Figure 3 show that an increase in the concentration of HCl had a progressive positive effect on the leaching dissolution of Li before starting to plateau, while a parabolic trend is seen for Fe dissolution. Again, the leaching dissolution efficiency of Li increased with increasing H₂O₂, achieving its maximum at the highest dosage studied. Conversely, the addition of the oxidant shows a sharp decline in the iron co-dissolution, confirming its effective role in suppressing iron co-dissolution through oxidative precipitation.

Figure 3. The individual effect plots for Li extraction and Fe dissolution.

Furthermore, it can be observed that the solid-to-liquid ratio exerted the most significant impact on both responses. For lithium dissolution, a distinct parabolic curvature is visible, where extraction efficiencies peaked at intermediate solid-to-liquid ratios before declining sharply due to mass-transfer limitations and proton starvation. Similarly, increasing the S/L ratio continuously suppressed the iron co-dissolution, driving its dissolution down toward its minimum.

The two-dimensional (2D) contour plots in Figure 4 show the combined effects of the significant input parameters on Li and Fe leaching dissolution efficiency. Each plot depicts the interaction between two input parameters while holding the third constant. For the Li contour plots, the results show that the interaction between H₂O₂ and HCl concentration had an insignificant effect on Li dissolution within the range studied. In contrast, Li extraction was significantly influenced by the interaction between the S/L ratio and HCl concentration. It can be observed that increasing the HCl concentration positively affected Li extraction as the S/L ratio increased from 10 to ≈35%. Similarly, the interaction between the S/L ratio and H₂O₂ had a significant effect on Li extraction, with Li extraction increasing with H₂O₂ across lower and intermediate S/L ratios. There is a sharp drop in the dissolution efficiency when high S/L ratios are combined with low H₂O₂ dosages. Furthermore, the Fe contour plots are also shown in Figure 4.

Figure 4. Two-dimensional contour plots showing the effects of the individual parameters on Li extraction and Fe dissolution. The contour color gradients represent the dissolution efficiency (%), with darker color shades denoting higher percentage values.

The results clearly show that the interaction between the H₂O₂ dosage and HCl significantly influenced Fe dissolution within the range studied. It can be observed that increasing the H₂O₂ dosage minimized iron co-dissolution between 2 and 2.5 M HCl. In contrast, the interaction between the S/L ratio and HCl had a relatively insignificant effect on iron leaching. Lastly, the interaction between the S/L ratio and H₂O₂ dosage had a prominent impact on Fe leaching. The statistical analysis confirmed that the developed response surface model adequately describes the selective leaching behavior of the LFP black mass and can therefore be used to evaluate the influence of the operating variables on lithium extraction.

3.2. Parameter Sensitivity and Physical Chemistry Mechanisms

To further interpret the statistically validated response surface model and identify the operating conditions that maximize lithium extraction, the influence of each process variable was examined individually using parameter sensitivity plots generated from the 17-run Box–Behnken Design (BBD) experiments. Although these plots illustrate the effect of varying one factor while the remaining variables are held at their central levels, they are derived from the complete RSM dataset rather than from separate one-factor-at-a-time experiments. The influence of each process parameter is discussed in the following sections. Figure 5 displays the effects of the various factors on the selective leaching of Li from LFP black mass.

Figure 5. Effect of (a) solid-to-liquid ratio, (b) HCl concentration, (c) H₂O₂ dosage on the selective leaching of Li.

3.2.1. Effect of Solid-to-Liquid Ratio on LFP Selective Leaching

The effect of the S/L ratio on dissolution efficiency and the Li-Fe selectivity is shown in Figure 5a. This evaluation was done by isolating specific response vectors from the BBD matrix, where the S/L ratio was increased from 10% to 50%, while alternative parameters were fixed at their statistical center points (2.0 M HCl and 2% v/v H₂O₂) at room temperature. When the S/L ratio was increased from 10% to 50%, there was a significant decrease in Li extraction (from 86% to 43%), accompanied by a concurrent rise in selectivity (from 53 to 73). These findings demonstrate a strong alignment with previous studies on LFP leaching, while also highlighting critical benchmarks. The observed negative correlation between the S/L ratio and dissolution efficiency is consistent with the results of refs. [48,49]. In their work, Li extraction typically decreased at high S/L ratios. This was attributed to the reduced contact between the LFP black mass and the leaching agent, which restricts the diffusion of ions at the solid–liquid interface. It can be inferred that at lower solid loadings (10%), HCl is in relative excess compared to the available solid mass, providing an abundant chemical driving force for near-complete dissolution. The stoichiometric relationship between HCl consumption and the LiFePO₄ dissolution can be established from the primary selective delithiation reaction (Equation (7)):

$$2LiFeP{O}_{4\left(s\right)}+2HC{l}_{\left(aq\right)}+H_2{O}_{2\left(aq\right)}\to 2Li{Cl}_{\left(aq\right)}+2FeP{O}_{4\left(s\right)}+2H_2{O}_{\left(l\right)}$$

(7)

From Equation (7), the required acid-to-mineral molar ratio is 1:1. Given an ICP-OES head grade of 32.12 wt% Fe (translating to 90.67 wt% active LiFePO₄), a 100 mL volume of 2.0 M HCl supplies 0.20 moles of H⁺ against a molar demand of 0.06 moles for a 10.0 g solid mass. This yields an acid excess factor (λ) of 3.48, providing a steep concentration gradient across the fluid boundary layer. Conversely, increasing the S/L ratio to 50% w/v expands the solid mass to 50.0 g (0.29 moles LiFePO₄), compressing λ to 0.70. This establishes an absolute theoretical acid deficit of 30.4%, mathematically capping the maximum possible Li extraction at 69.6% and confirming that localized proton starvation limits the system. This confirms that the sharp extraction decline at 50% w/v results from a coupled physicochemical regime: stoichiometric proton deficiency ($\lambda =0.70$) caps the maximum theoretical yield, while elevated slurry viscosity drastically impairs hydrodynamic shear, thickening the stagnant Nernst boundary layer around the particles and suppressing both oxidant diffusion and outward Li⁺ mass transport.

3.2.2. Effect of HCl Concentration on LFP Selective Leaching

The impact of increased acid strength on selectivity, Li and Fe dissolution efficiency is depicted in Figure 5b. When HCl concentration was increased from 2.0 M to 3.0 M (at a constant 3% H₂O₂ dosage and S/L ratio of 30%), the Li dissolution increased slightly from 86% to 92% but caused Fe co-dissolution to more than triple (from 13% to 44%) and, consequently, dropped selectivity significantly from 88 to 72. This observation aligns with the selectivity threshold noted by ref. [49] and ref. [37]. While H⁺ ions are critical for disrupting the stable olivine LFP lattice, excessive acid strength also impedes the reprecipitation of FePO₄·2H₂O, thereby increasing impurity levels in the leachate. According to the shrinking core model utilized by ref. [46], leaching is limited by the internal diffusion of acid through an FePO₄ product layer. Surpassing stoichiometric requirements leads to the chemical dissolution of this protective FePO₄ layer, shifting the process from selective leaching to total leaching. This breakdown is supported by the Eh-pH diagram shown in Figure 6. The diagram demonstrates that while solid FePO₄·2H₂O remains stable under moderately acidic, oxidative conditions, transitioning the system into a highly concentrated acid region (pH ≤ −3) completely suppresses the solid phase, shifting the thermodynamic equilibrium into the entirely soluble, aqueous H₃PO₄, Fe³⁺, Li⁺ region. Under these highly acidic boundary conditions, the protective layer undergoes chemical breakdown into Fe³⁺ species. This breakdown is further supported by Mohammed Zadeh et al. [40], who observed that at low pH ranges, dissolution of Fe and P can reach nearly 50%, which mirrors the selectivity loss observed at 3.0 M HCl. Thus, an optimal leaching environment must balance proton activity and redox potential to achieve selective Li dissolution.

Figure 6. The E-pH diagram for the Li-Fe-P-H₂O system at 25 °C [53].

3.2.3. Effect of H₂O₂ Dosage on LFP Selective Leaching

To evaluate the impact of H₂O₂ dosage on leaching selectivity and Li and Fe dissolution efficiency, the H₂O₂ dosage was varied while maintaining a fixed S/L ratio of 30%, 2.0 M HCl, a 60 min leaching duration, and room temperature (Figure 5c). At 1% H₂O₂ dosage, iron leaching is significant (75%), resulting in low selectivity (S_Li = 58%). Increasing the dosage to 3% suppresses iron co-dissolution to 13%, while maintaining stable Li dissolution (86%) and surging the selectivity upwards to S_Li = 88%. This phenomenon can be explained as H₂O₂ oxidizes Fe²⁺ to Fe³⁺, which combines with PO₄³⁻ to form FePO₄, making Li more easily leached. This performance is in consensus with the results of ref. [46] and ref. [37]. Zhang et al. [46] found that in the absence of H₂O₂, selectivity was only 66.93%, but it surged significantly upon the addition of H₂O₂. Similarly, Cong et al. [37] also noted that raising H₂O₂/Li ratios to an optimal ratio of 2.13 effectively increased Li extraction while simultaneously decreasing Fe co-dissolution. This observed phenomenon may be justified by the iron-phosphorus fixation mechanism described by ref. [50], who identified H₂O₂ as the primary reagent for ensuring Fe remains in the solid phase while Li⁺ is released. This aligns with the in situ oxidation model, where H₂O₂ elevates the solution oxidation–reduction potential, theoretically poising the system within the thermodynamic stability domain of solid FePO₄·2H₂O (>0.60 V vs. SHE, as delineated in the Eh-pH diagram in Figure 6). In the absence of an oxidant (non-oxidative acid leaching), HCl promotes congruent, unselective dissolution of both Fe²⁺ and Li⁺ into solution. Introducing H₂O₂ triggers topotactic oxidative extraction: Fe²⁺ is oxidized to Fe³⁺ within the olivine framework, prompting the ejection of Li⁺ into the aqueous phase while retaining the three-dimensional phosphate framework as insoluble solid-state FePO₄. While direct morphological and phase characterization of the washed residue (XRD/SEM-EDS) will further quantify crystal lattice parameters in larger pilot demonstrations, the elemental mass balance confirms this mechanism, with over 80% of the total iron inventory immobilized in the leach residue. Despite H₂O₂ having the potential to achieve high metal separation, at higher oxidant dosages, it levels off Li dissolution. Ref. [46] and ref. [37] stated that, once the H₂O₂/Li molar ratio reaches a specific threshold (~2.1 to 2.5), Fe²⁺ is completely oxidized, and further oxidant additions yield diminishing returns. Furthermore, excessive H₂O₂ concentrations are prone to catalytic decomposition into water and oxygen, particularly in the presence of Fe³⁺.

3.3. Optimization and Model Experimental Validation

Based on the results obtained under varying leaching conditions, the leaching behavior of Li and Fe highlights a significant tradeoff. While Li is easily liberated through acid-driven dissolution, Fe suppression relies on maintaining oxidative conditions by H₂O₂ addition to induce precipitation as solid FePO₄. To resolve this, the developed second-order regression models were utilized to estimate optimum operating conditions using the Minitab software desirability approach. In Table 8, recommendations for Li and Fe dissolution efficiencies (or extraction efficiencies) are provided when leaching EoL LFP batteries: 32.86% (S/L ratio); 2.34 M HCl and 3.0 vol% H₂O₂ at room temperature for 60 min will generate the highest trade-off. Verification of the model was done again in the laboratory to look at the suitability of the experimental results with the model obtained.

Table 8. Experimental results under optimal process parameters.

Parameters	Optimum Condition	Value of Response	Predicted	Experiment	Error
Solid–Liquid Ratio (w/v%)	32.86	Li extraction (%)	86.34	86	0.39
HCl Concentration (M)	2.34	Fe dissolution (%)	20	19.20	4.02
H2O2 Dosage (vol%)	3.00

Experimental verification was carried out in triplicate ($n=3$) at the optimal condition (32.86% w/v solid-to-liquid ratio, 2.34 M HCl, and 3.0 vol% H₂O₂), resulting in mean Li and Fe dissolution efficiencies of 86.0 ± 1.2% and 19.2 ± 0.8%, respectively. These closely match the model-predicted values of 86.34% for Li extraction and 20.0% for Fe co-dissolution, demonstrating low percentage relative errors of 0.39% and 4.02%, respectively. In addition, the five center-point replicate runs in the BBD matrix demonstrated high experimental precision, with standard deviations of ±3.6% for Li (mean 85.2%) and ±1.6% for Fe (mean 31.0%), confirming robust experimental repeatability and validating the low residual pure error in ANOVA. In other words, Li is highly dissolved, and Fe is successfully suppressed as graphically illustrated in Figure 7.

Figure 7. Comparison of predicted and experimental validation values for the Li extraction and Fe dissolution under optimized conditions.

3.4. Purification of LFP Leachates

When studying iron removal from the HCl-H₂O₂ LFP-based PLS, the precipitation percentage at 60 °C for Fe was found to be high (99.0%), even at the lowest studied pH level of 2.5, as shown in Figure 8.

Figure 8. Precipitation percentages of Fe and Li from the HCl-H₂O₂ based LFP leachate (PFe* and PLi* represent the percentage of Fe precipitated and Li co-precipitated, respectively).

When 3 M NaOH was used to modify the pH, the precipitation percentage increased to a near-complete value of 99.7% at pH 3.0 and reached a plateau at 99.8% at a pH of 3.5, as presented in Table 9.

Table 9. Fe and Li precipitation behavior for the H₂O₂-NaOH LFP precipitation system.

Sample Code	Initial pH	Target pH	PFe*	PLi*
S2202	0.35	2.5	99.0	6.8
S2203	0.35	3	99.7	0.6
S2204	0.35	3.5	99.8	0.4
S2179	0.35	4	46.9	10.6

PFe* and PLi* represent the percentage of Fe precipitated and Li co-precipitated, respectively.

Since substantial Fe precipitation was supported by the availability of PO₄³⁻ ions, this finding demonstrates that optimum separation efficiency between Li and FePO₄ may be attained for LFP-based leach solution at pH 3 to 3.5. This behavior aligns closely with findings by Chang et al. [54], who identified pH 3 as the optimal equilibrium point for achieving high-purity FePO₄ with a precipitation percentage of 99.57%. Similarly, Zou et al. [55] reported a high precipitation percentage of 97.8% at pH 2.0, rising to > 99% for pH levels between 2.5 and 4.5. In the current study, the use of 1% vol. of H₂O₂ (w/w%), equivalent to 3.7 mL, for pretreatment ensured sufficient oxidation of Fe²⁺ to Fe³⁺, shifting the equilibrium toward the stable FePO₄ solid phase. This oxidation step is essential for successful precipitation and follows Equation (8):

$$2{Fe}_{\left(aq\right)}^{2+}+H_2{O}_{2\left(aq\right)}+2H_{\left(aq\right)}^{+}\to {2Fe}_{\left(aq\right)}^{3+}+2H_2{O}_{\left(l\right)}$$

(8)

The high Fe precipitation percentage at low pH is primarily attributed to the availability of PO₄³⁻ ions, which facilitates the formation of amorphous FePO₄ via Equation (9):

$${Fe}_{\left(aq\right)}^{3+}+P{O}_{4\left(aq\right)}^{3-}\to FeP{O}_{4\left(s\right)}$$

(9)

As illustrated in the Pourbaix (Eh vs. pH) diagram for the Li-Fe-P-H₂O system (Figure 6), Fe³⁺ preferentially precipitates as FePO₄∙2H₂O at acidic pH levels (typically pH 2.0–3.0) when the redox potential is maintained above 0.5 V. This theoretical stability region directly supports the near-complete precipitation observed in the current study’s pH 3.0–3.5 window. To further evaluate the competition between phosphate and hydroxide species, a speciation diagram was simulated using Hydra-Medusa software (32-bit version, KTH Royal Institute of Technology, Stockholm, Sweden) as shown in Figure 9. The simulation confirms that under the experimental conditions, FePO₄(s) is the sole dominant species up to ≈ pH 5.0. This theoretical dominance confirms that the iron removed in the pH 3.0–3.5 window is recovered primarily as a phosphate salt rather than a hydroxide, ensuring high product purity as justified by the Ksp values reported by [55].

Figure 9. Speciation diagram for Fe-PO₄³⁻ by Medusa and Hydra software.

In their studies, the authors of [55] further justify this by noting that FePO₄ has a significantly lower solubility product constant (Ksp = 1.4 × 10⁻²⁷) compared to AlPO₄ (Ksp = 4.0 × 10⁻²⁰), ensuring that iron phosphate precipitation is favored over other species, including hydroxides, in a pH range of ~ 3.0. In contrast to the high precipitation percentages observed at mildly acidic levels (pH 3.0–3.5), an apparent precipitation drop was recorded at pH 4.0 (sample S2179), where the calculated Fe removal efficiency decreased to 46.9% while Li loss increased to 10.6%. While thermodynamic speciation (Figure 9) predicts sustained stability of solid FePO₄ and insoluble Fe(OH)₃ above pH 3.5, this empirical outcome is governed by solid–liquid separation and colloidal phenomena rather than thermodynamic dissolution. Rapid dropwise addition of concentrated 3.0 M NaOH to achieve pH 4.0 creates intense localized basicity zones, inducing burst nucleation of amorphous, non-settling colloidal ferric oxyhydroxide and hydrated hydroxide precipitates according to Equation (10):

$${Fe}_{\left(aq\right)}^{3+}+3O{H}_{\left(aq\right)}^{-}\to Fe(OH{)}_{3\left(s\right)}$$

(10)

These sub-micron colloidal fines resist standard gravitational settling and partially penetrate or bypass standard filtration media. When the resulting PLS filtrate is stabilized with dilute acid prior to ICP-OES analysis, these entrained colloidal particulates redissolve, registering as apparent aqueous iron in solution. Furthermore, the elevated viscosity and high specific surface area of this gelatinous, uncoagulated phase substantially exacerbate mechanical liquor retention, directly accounting for the observed spike in lithium entrapment (10.6%). Consequently, operating at a controlled equilibrium window of pH 3.0–3.5 represents the critical process boundary, balancing rapid crystalline macro-floc aggregation, minimal colloidal peptization, and negligible lithium entrainment (<0.6%).

3.5. Synthesis and Purity Evaluation of Lithium Carbonate

Following oxidative precipitation, the pregnant leach solution (PLS) still contained residual iron at a level unsuitable for high-purity Li₂CO₃ production. For instance, sample S2203 (initial pH 3.0) had an iron concentration of 19 mg/L. To achieve further iron removal, an oxidative polishing step was performed on the Li-rich solution at a target pH of 3.5 using 2 mL of 30 wt.% H₂O₂ and 1 M NaOH. This trial was designated S2231. During polishing, H₂O₂ addition ensured complete oxidation of residual Fe²⁺ to Fe³⁺, resulting in quantitative precipitation of iron as Fe(OH)₃. The solution volume increased from 70 mL to 100 mL (specific volume expansion factor = 1.43). Raising the pH from 3.0 to 3.5 reduced the iron concentration below the ICP detection limit (ND), corresponding to >99% iron removal. Accounting for the dilution factor from volume expansion (70 mL to 100 mL), the absolute mass of lithium decreased from 356.0 mg (5085 mg/L in 70 mL) to 245.9 mg (2459 mg/L in 100 mL), representing an actual lithium mass loss of 30.9% (approximately 31%, rather than the 51.6% concentration decrease). This loss is primarily attributed to physical liquor entrainment within the voluminous, gelatinous Fe(OH)₃ precipitate, colloidal surface adsorption, and localized sampling variances during slurry separation. In commercial continuous circuits, this entrained lithium can be mitigated and largely recovered via multi-stage counter-current displacement cake washing, typically requiring 1.5–2.5 L wash water per kg of dry cake (wash ratio 1.5–2.5), and by recycling the resulting wash filtrates back to the primary leaching circuit to maintain high cumulative recovery. Subsequent lithium precipitation from the polished PLS (S2231) was performed by addition of Na₂CO₃. This reagent was selected because lithium chloride is difficult to dry at industrial scale, whereas Li₂CO₃ is readily handled as a solid product. Precipitation from dilute solutions (Li < 10 g/L) is known to be sensitive to temperature, CO₃²⁻/Li⁺ molar ratio, and ionic strength. Literature indicates that efficient carbonate precipitation typically requires an initial lithium concentration of at least 7.5 g/L [51]. Since the polished PLS contained only 2.459 g/L Li, the solution was preconcentrated by a factor of 5.0 prior to precipitation. This yielded a single-pass lithium precipitation efficiency of 29.7% (73 mg Li precipitated), leaving a residual concentration of 1572 mg/L in the filtrate (S2256). This moderate single-stage recovery is governed by the retrograde solubility of Li₂CO₃ (~7.2 g/L at 90 °C) and the remaining aqueous volume. In commercial recycling plants, overall lithium recovery is closed-loop: the mother liquor is continuously recycled back to the thermal evaporation circuit to utilize residual lithium ions and exploit the common-ion effect (${\mathrm{CO}}_3^{2-}$ excess), driving cumulative circuit recovery above 85–90%.

Based on the aqueous Li⁺ mass balance from ICP-OES measurements of solution depletion, the precipitation of 73 mg Li stoichiometrically equates to a theoretical pure Li₂CO₃ mass of 0.39 g. Comparing this value to the actual weighed dry cake mass of 0.40 g yielded an estimated chemical purity of 97.25% (≈ 97.3%) for the recovered crude product. However, as noted from the moderate single-stage precipitation efficiency (29.7%), this indirect calculation does not exclude the co-precipitation or physical inclusion of entrained soluble salts, primarily NaCl and trace sodium/iron species. The post-precipitation filtrate contained 10 mg/L residual iron, likely due to increased alkalinity after Na₂CO₃ addition and stabilization of colloidal iron species. While the recovered solid represents a technical-grade lithium precursor, upgrading to certified battery-grade specifications (>99.5% Li₂CO₃) under industrial flowsheets requires subsequent secondary purification, such as carbonation into soluble lithium bicarbonate (LiHCO₃), polishing filtration to reject insoluble silicates/hydroxides, and thermal decomposition recrystallization.

The crystallographic phase purity and structural characteristics of the precipitated product were directly evaluated via powder X-ray diffraction (Figure 10). As shown, all predominant diffraction reflections align with the monoclinic zabuyelite phase (Li₂CO₃, PDF 00-022-1141) with distinct diffraction peaks indicative of high crystallinity. Minor baseline reflections correspond to trace reference/substrate markers (PDF 00-072-0420) without detectable co-precipitated secondary crystalline lithium salts, supporting the structural integrity of the recovered technical-grade product.

Figure 10. Powder X-ray diffraction (XRD) pattern of the synthesized Li₂CO₃ precipitate (black line), indexed against the standard reference pattern for zabuyelite (Li₂CO₃, PDF 00-022-1141; red lines) and Al₂O₃ (PDF 00-072-0420; blue lines). Data collected using Cu–Kα radiation (λ = 1.54060 Å).

3.6. Integrated Flowsheet

Based on the data obtained, a flowsheet for the selective recovery of lithium as Li₂CO₃ using the HCl–H₂O₂ system from the LFP black mass is presented in Figure 11. The process begins with transferring the valuables in the LFP black mass via a selective leaching step. The black mass is contacted with an optimized lixiviant consisting of 2.3 M HCl and 3 wt% H₂O₂ at a high S/L ratio of 33% (w/v). Under these conditions, lithium is preferentially solubilized, while a substantial fraction of the iron remains in the solid phase as an insoluble ferric phosphate (FePO₄). This results in an optimized pregnant leach solution (PLS) containing 86% of the available lithium and only 19.2% of the iron. Following solid–liquid separation, impurities such as residual iron in the PLS are removed in a purification stage using an oxidative NaOH-H₂O₂ process, precipitating the iron as a mixed solid phase of FePO₄ and insoluble Fe(OH)₃ (as outlined in Section 3.4). To ensure complete removal of any remaining multivalent transition metal impurities before lithium recovery, the solution undergoes an oxidative polishing step with controlled H₂O₂ additions. The purified PLS is then subjected to thermal preconcentration, which simultaneously increases the lithium concentration and induces fractional crystallization of sodium chloride (NaCl) salts. Finally, lithium is recovered by adding a saturated Na₂CO₃ solution at elevated temperature, yielding a 97.25% Li₂CO₃ purity. This integrated sequence demonstrates effective lithium recovery with minimal iron crossover, providing a low-energy, selective route for valorising LFP black mass.

Figure 11. Proposed flowsheet for the selective recovery and purification of Li from LFP black mass using HCl–H₂O₂ lixiviant system.

3.7. Scalability Assessment and Techno-Economic Considerations

To evaluate the industrial scalability and practical impact of this process, the operating parameters identified in this work (32.86% w/v solids loading at 25 °C) were compared against conventional benchmark hydrometallurgical systems (typically 50–100 g/L solids loading at 60–90 °C). Intensifying the pulp density to ~33% w/v yields an immediate threefold reduction in required reactor working volumes and liquid effluent handling for an equivalent annual throughput of spent LFP black mass. This volumetric compression reduces equipment sizing, agitation vessel footprints, and facility capital expenditure (CAPEX). From an operational expenditure (OPEX) standpoint, conducting selective delithiation at ambient room temperature (25 ± 2 °C) eliminates the substantial thermal energy penalty associated with heating large volumes of acidic slurry to 60–90 °C. Furthermore, leaching at high solids loadings directly generates a concentrated pregnant leach solution (>5.0 g/L Li), substantially lowering the thermal duty required for downstream evaporative preconcentration prior to lithium carbonate precipitation. While operation at 32.86% w/v offers substantial capital and thermal savings, scale-up introduces preliminary operational risks that require targeted engineering mitigations: (i) Slurry Rheology and Agitator Stalling: High solids loading induces non-Newtonian shear-thinning behavior, which risks localized dead zones and motor overload; this is mitigated through high-torque, dual-impeller hydrofoil systems (e.g., Lightnin A310/A320) optimized for uniform off-bottom particle suspension. (ii) Exothermic Heat Accumulation: Peroxide decomposition and concentrated acid neutralization release exothermic enthalpy; in unheated large-scale reactors, internal cooling jackets or controlled staged addition of H₂O₂ are required to maintain slurry temperature near the optimal 25 °C boundary and prevent thermal runaway or parasitic H₂O₂ degradation. (iii) Cake Compressibility: Downstream filtration of fine-grained residues (d₅₀ ≈ 45 μm) poses blinding risks on filter cloths, which can be mitigated using continuous horizontal belt filters or automated filter presses equipped with high-pressure displacement washing. Incorporating these engineering controls ensures that the demonstrated thermal savings and reduced water circulation translate reliably to full-scale operations.

4. Conclusions

This current study investigated selective lithium leaching from LFP black mass using an intensified HCl–H₂O₂ lixiviant system over a range of solid-to-liquid ratios. Process parameters were optimized by response surface methodology based on a Box–Behnken design. Under the identified optimal conditions (2.34 M HCl, 3.0 vol.% H₂O₂, 32.86% w/v solids loading, 25 °C, 60 min), lithium extractionreached 86.0% while iron dissolution was limited to 19.2%, effectively retaining the iron-phosphate matrix in the solid residue.

Downstream purification of the pregnant leach solution was performed by addition of 3.0 M NaOH at 60 °C for 30 min within a controlled pH window of 3.0–3.5. This step precipitated 99.7% of the dissolved iron with minimal lithium co-precipitation (<0.55%). Subsequent oxidative polishing reduced residual iron to below detection limits, enabling recovery of lithium as technical-grade Li₂CO₃ with an estimated stoichiometric purity of 97.3%.

Future work could consider evaluating reagent recycling, effluent chloride management, and integration with industrial purification circuits to validate the route for commercial deployment.

5. Patents

The leaching process reported in this study adapts technology protected under European Patent Application No. EP4174196A1.