The Recycling of Plastics and Current Collector Foils from End-of-Life NMC-LCO Type Electric Vehicle Lithium-Ion Batteries Using Selective Froth Flotation

Abstract

The recent increase in end-of-life (EoL) lithium-ion batteries (LiBs) has become a significant concern worldwide. Most studies in the literature have primarily focused on recovering cathode active metals from black mass (BM), whereas the separation of anode–cathode foils, plastics, and casing metals which are the essential components of LiBs has received relatively little attention. To reduce costs and maximize the recovery of valuable metals in subsequent hydrometallurgical or pyrometallurgical processes, EoL LiBs require appropriate pre-treatment. This study aims to scrape off the BM adhering to the electrode foils resulting from gradual crushing and subsequently separate the plastics and copper (Cu) from other metals through a two-step selective flotation process. The results demonstrated that plastics, due to their natural hydrophobicity, could be effectively removed using a frother, achieving more than 95% recovery with less than 5% metallic contamination. Following plastic flotation, Cu particles were floated in the presence of 3418A, yielding a Cu concentrate containing 65.13% Cu with a recovery rate of 96.4%. Additionally, the aluminum (Al) content in the non-floating material, remaining in the cell, increased to approximately 77%.

1. Introduction

Developments in energy storage technologies have brought significant transformations across various sectors and have become crucial for achieving a sustainable future. Moreover, the rising energy demand, increased use of renewable energy sources, and extensive adoption of electric vehicle (EV) technologies have contributed to the need for advanced energy storage systems. In particular, high fossil fuel consumption and the necessity to reduce environmental damage have driven efforts to develop renewable and sustainable energy sources such as high-quality batteries, wind turbines, fuel cells, and solar panels [1,2,3].

LiBs are characterized by their compact size, lightweight design, high energy density and voltage, extended lifespan, absence of memory effect, efficient operation across a wide temperature range, and eco-friendly properties [1,4]. The growing use of EVs and consumer electronics has significantly increased the demand for LiBs, resulting in a substantial rise in the volume of EoL batteries requiring disposal [5]. For example, China produced approximately 15 billion LiBs in 2014, and by 2024, its annual production capacity surpassed 1170 GWh, representing around 76% of global output [6]. This surge in LiB demand highlights their critical role in energy storage, EV charging infrastructure, and large-scale battery energy storage systems (BESS). Consequently, the projected increase in LiB production is expected to lead to a significant volume of EoL batteries.

Incorrect disposal of spent LiBs poses severe environmental problems, including soil and water contamination from leakage of heavy metals and hazardous compounds. Beyond environmental and economic benefits, exploiting the resource potential of battery scraps is essential for conserving natural resources and promoting the sustainable development of metals and related industries [7,8]. Additionally, EoL LiBs contain many components with high economic value, such as lithium (Li), cobalt (Co), copper (Cu), aluminum (Al), nickel (Ni), iron (Fe), manganese (Mn), and graphite [9,10].

LiBs are manufactured to meet specific application requirements, with the most variation occurring in the active cathode material. Several types of LiBs have been developed, including lithium cobalt oxide (LCO), nickel cobalt manganese oxide (NMC), nickel cobalt aluminum oxide (NCA), lithium manganese oxide (LMO), lithium iron phosphate (LFP), and lithium titanate oxide (LTO). Among these, NMC, LCO, LFP, and LMO batteries are the most widely used in commercial applications [11]. Figure 1 illustrates the main components of a typical LiB cell, including the cathode and anode coatings (BM), current collector foils, separator, electrolyte, and casing. NMC-based LiBs generally comprise 5%–20% Co, 5%–7% Li, 7%–10% Cu, 10%–15% Al, 4%–5% Ni, 4%–5% Mn, around 15% organic compounds, and approximately 7% plastic materials, indicating considerable potential for resource recovery [12,13]. Therefore, EoL LiBs are increasingly regarded as strategic secondary raw materials. Recycling these batteries enables the reclamation of critical metals, thus promoting sustainable resource management and contributing to the circular economy [14,15].

Driven by technological advancements, increasing market demand, and the growing volume of waste batteries, global battery recycling capacity has expanded from 9 GWh in 2019 to 50 GWh in 2024, and it is projected to exceed 230 GWh by 2030 [16]. Recycling LiBs facilitates the recovery of critical metals, reducing dependency on mining, conserving natural resources, and supporting sustainable production models. It is estimated that recycling 500,000 tons of LiBs could recover approximately 45,000 tons of Cu, 15,000 tons of Al, 90,000 tons of Fe, 75,000 tons of Li, 60,000 tons of Co, and 35,000 tons of plastic materials [17].

The recycling process of LiBs typically begins with a series of pre-treatment steps designed to prepare dismantled EoL battery packs for main recovery. These steps are crucial for purifying core components such as cathode and anode materials, enhancing the efficiency and safety of subsequent processing [18]. The primary aim of pre-treatment is to achieve effective and safe separation of battery components, which is particularly important for hydrometallurgical recovery and direct recycling methods [19]. Due to the structural complexity and hazardous nature of LiBs, pre-treatment is essential to mitigate safety risks and facilitate the selective recovery of valuable materials. Standard pre-treatment operations include deep discharging to eliminate residual charge, mechanical size reduction, particle classification, physical component separation, thermal and chemical treatments, and cathode material isolation [19,20,21]. This stage establishes foundational conditions to ensure high recovery yields and environmental compliance in advanced recycling technologies.

Size reduction is crucial for effective liberation of electrode materials and directly influences the efficiency of subsequent purification and enrichment stages. Shredders, hammer mills, and impact crushers are commonly used to break down battery cells into smaller fragments [8,14,22,23]. Further size reduction, if needed, can be performed by grinding into a fine powder. Screening systems then classify materials by particle size, separating fine fractions, rich in Li, Co, and graphite, from coarser fractions containing plastics, Fe, Cu, and Al. The fine fraction (undersize) comprises valuable BM suitable for hydrometallurgical or direct recovery, while the coarse fraction (oversize) can be further processed or recycled, depending on its composition.

The primary recycling technologies for LiBs include pyrometallurgical, hydrometallurgical, and direct recovery methods [24,25]. Pyrometallurgical processes operate at high temperatures, resulting in high energy consumption and the emission of harmful gases [26]. Hydrometallurgical methods are favored for lower energy use, higher recovery efficiency, and reduced environmental impact [27]. These acid leaching processes typically use sulfuric acid, with recent research focusing on replacing traditional reducing agents like hydrogen peroxide with environmentally friendly organic compounds such as glucose, tea waste, and grape seed extracts [2]. Despite their benefits, these methods often involve complex chemical steps, high energy demands, or limited recovery yields.

Flotation separates materials based on their hydrophobic (water-repelling) and hydrophilic (water-attracting) properties by causing some particles to float while others sink. Flotation of coarse metallic and plastic fractions from EoL LiBs shares many mechanistic similarities with classical ore beneficiation. Recent work by Aarab et al. [28] on the eco-friendly flotation of apatite using bio-based collectors demonstrates the potential of sustainable reagent design and adsorption mechanisms that are also relevant to plastic–metal separation. Originally used for ore beneficiation, flotation has recently been applied effectively for the direct recycling of EoL LiBs [29]. This method is advantageous due to its low energy consumption, minimal chemical requirements, and ability to separate graphite from cathode metal oxides based on surface characteristics. The naturally hydrophobic graphite anode facilitates the separation of carbon-based materials in BM [30]. In flotation, graphite particles attach to bubbles and float, while mostly hydrophilic cathode materials like LCO and NMC sink to the bottom fraction [31].

Despite growing interest in LiB recycling, studies focusing on the selective recovery of coarse metallic fractions such as Cu, Al, Fe, and Ni are scarce. Saneie et al. [32] demonstrated effective separation of Cu and Al foils in the −4 + 1 mm size range from 18,650-type spent LiBs using a two-stage flotation process involving initial plastic removal followed by Cu flotation. To our knowledge, only a limited number of studies have reported flotation-based separations relevant to spent LiBs; most target electrode powders/graphite or small-device batteries rather than EV-grade mixed NMC/LCO coarse fractions. Prior research on flotation-based recovery from waste printed circuit boards (PCBs) [33,34,35] suggests that liberated plastic particles can be effectively separated from each other and metallic components in coarse fractions using froth flotation, indicating promising potential for LiB recycling.

Accordingly, this study focuses on the selective separation of plastics, Cu, Al, and structural metals from NMC and LCO-type LiBs through the application of various surface-active reagents. After initial BM separation from the metallic fraction via sequential size reduction and sieving across four stages, the first flotation step targeted the selective removal of naturally hydrophobic plastics using a frother (Methyl Isobutyl Carbinol, MIBC). In the subsequent flotation stage, the effect of different collectors on Cu recovery was examined by promoting hydrophobic Cu particles to the froth phase, allowing hydrophilic Al and structural metal particles to remain in the pulp.

2. Experimental Studies

2.1. Materials

In this study, approximately 60 kg of mixed NMC and LCO-type EoL EV LiBs were supplied by Exitcom Co., located in Kocaeli, Türkiye. A multistage size reduction and classification system was applied, as illustrated in Figure 2. Initially, the batteries were fully discharged for 48 h using a 5% NaCl solution, and then air-dried. Following this, primary crushing was performed using a dual-shaft shredder at the recycling company to ensure safety. The crushed material (particles smaller than 40 mm) was then transferred to the Mineral Processing Engineering Department’s Prof. Dr. Güven Önal Pilot Plant at Istanbul Technical University for subsequent comminution processes.

The BM 1 sample was initially separated using a 0.2 mm sieve because this cut size efficiently removed most of the liberated BM while retaining metallic foils and plastics for subsequent flotation. This choice was based on preliminary screening tests and is independent of the crusher settings. The coarse fraction was subsequently reduced to particles smaller than 10 mm using a shredder. The crushed material was screened again with a 0.2 mm sieve, and the oversized fraction was fed into a four-blade cutting mill (RAM200 model, supplied by RANTEK Co., Ankara, Türkiye) to further scrape off the residual BM adhering to the current collectors. This cutting mill is designed to uniformly crush elastic, soft, medium-hard, fibrous, and heterogeneous materials. The shear forces reduce particle size while maintaining the flat surface shape of the material, which is advantageous for the flotation process. Particle size adjustment was achieved by using interchangeable sieves (6, 4, 2 mm, etc.) located in the lower compartment of the mill. The BM adhering to the surfaces of the current collectors was further collected using a 0.2 mm sieve and preserved for future studies.

The particle size range of −2 + 0.2 mm was selected based on preliminary sieving tests and chemical analysis, which indicated that this fraction contained sufficiently liberated plastic and metal particles for efficient separation. The finer fraction (<0.2 mm) contained most of the BM powders and electrode coatings, which are more suitable for chemical leaching rather than flotation. Although flotation is conventionally applied to particles below ~0.5 mm due to buoyancy limitations of larger spherical particles, the plate-like shape of electrode foils and plastic pieces in this waste material enabled successful flotation at coarser sizes. Additional tests conducted with the −4 + 0.2 mm fraction resulted in poor separation efficiency due to excessive residual active material and insufficient liberation, confirming −2 + 0.2 mm as the optimal size range for flotation.

The coarse fraction, constituting approximately 32% of the feed material, mainly consisted of Cu foils, Al foils, battery casings, plastics, and separators, and was used in characterization and flotation experiments. KAX (Potassium Amyl Xanthate, >98%) (ECS chemicals, Türkiye) and Aerophine-3418A, Aero-3739, Aerofloat-242, Aerofloat-211 (Syensqo, Türkiye), and MIBC (ECS chemicals, Türkiye) were employed as collectors and frother, respectively, in the flotation tests. All flotation tests were performed using municipal tap water from the Terkos system (Istanbul), with a conductivity of approximately 620 μS/cm and pH around 7.5. The main dissolved species were Ca²⁺ (≈85 mg/L), Mg²⁺ (≈12 mg/L), Na⁺ (≈28 mg/L), Cl⁻ (≈30 mg/L), and HCO₃⁻ (≈230 mg/L). The natural pulp pH during flotation was 8.3 ± 0.1 and remained stable throughout the experiments. Batteries were safely discharged in 5 wt.% NaCl solution according to widely accepted recycling practice. Although residual chlorides could potentially adsorb onto current collector surfaces, the subsequent comminution and water-based slurry preparation steps are expected to remove most soluble salts before flotation.

After quartering the comminuted material, a representative sample was collected, and sieve analysis was conducted using an electric shaker with standard sieves. The experimental data were interpreted based on product content and recovery. The compositions of major and trace elements were studied via a BRUKER S8 TIGER model wavelength dispersive X-ray fluorescence spectrometer (WDXRF) at the Geochemistry Research Laboratory of the ITU Department of Geological Engineering. For this research, milled samples were homogenized with wax at a 5:1 (w/w) ratio and pelletized using a HERZOG model pelletizer to create homogeneous discs suitable for XRF measurements. The accuracy of the analytical method was evaluated by analyzing the certified reference material, CRM (USGS DNC-1A and GEOSTATS GBM915-4), under the same sample preparation and measurement conditions. Three replicate analyses were performed, and the mean values were compared with the certified and/or reference values; the measured values showed good agreement with the certificate. The detection limits are 100 ppm for Al, Mn, and Fe, and 4.7, 2.2, and 1.5 ppm for Cu, Co, and Ni, respectively. The Loss on Ignition (LOI) study was performed in two consecutive stages to ascertain the volatile content of the samples. Initially, 2–3 g of material was combusted at 500 °C for 2 h to quantify organic matter and polymeric binders, including Polyvinylidene Fluoride (PVDF) and separator residues. Thereafter, the identical samples were combusted at 1000 °C for an additional 2 h to evaluate the total volatile content, encompassing carbonates and other thermally unstable constituents. This two-step LOI methodology facilitated a more thorough assessment of both organic and inorganic volatiles, especially in the BM and flotation-derived products. This approach is commonly reported for battery recycling studies to quantify Li₂CO₃ and other carbonates, which release CO₂ at temperatures > 800 °C. Although some volatilization of fluorides at 1000 °C is possible, we minimized exposure time to limit such effects.

XRF is constrained in its ability to detect light elements (Z < 11), including carbon (C), Li, oxygen (O), and fluorine (F), owing to the low-energy characteristics of their X-rays and matrix absorption phenomena [35,36]. Consequently, other methodologies were employed: lithium was measured via inductively coupled plasma (ICP) analysis using a PerkinElmer Avio 200 instrument, whereas C, O, and F were aggregated under the “others” group.

The recovery (R) used in result interpretation was calculated by Equation (1):

$$\mathrm{R} (\%) = (\mathrm{C}\mathrm{c}/\mathrm{F}\mathrm{f}) \times 100$$

(1)

where C is the weight of the concentrate, c is the metal content of the concentrate, F is the weight of the feed, and f is the metal content in the feed. All flotation tests were performed once under identical operating conditions, while the chemical assays for each product stream were conducted in triplicate to ensure analytical accuracy. In addition, the flotation experiment that yielded the best Cu recovery result (using 3418A) was repeated three times under the same conditions, and the results were consistent within experimental error.

To investigate the collector adsorption on Cu and Al electrode foils, a scanning electron microscopy (SEM) (Hitachi SU3500 model, Japan) coupled with an 80 mm² XMaxN silicon drift detector energy-dispersive spectroscopy (EDS) system (Oxford Instruments, Oxfordshire, UK). Cu collector electrode foils, whose surfaces were mostly cleaned of graphite by brushing from the anode side, and Al collector electrodes with cathode material still attached were used in the SEM studies. Sample preparation ensured that, after comminution and classification, the Cu foil surfaces were largely free of graphite due to the weak binding agent. In contrast, the cathode material is strongly adhered to the Al foil by the PVDF binder, making mechanical removal potentially damaging; therefore, Al samples were used without scraping. The foils were soaked for 10 min in 3000 g/t 3418A solution. After careful drying, they were cut into 5 × 5 mm pieces and mounted on aluminum stubs using double-sided carbon tape for SEM analysis. EDS results were interpreted qualitatively, as the quantification of light elements is known to be unreliable and susceptible to contamination. Therefore, the observed increases in surface C and S after collector conditioning are considered indicative of relative trends rather than absolute values. Both secondary electron (SE) and backscattered electron (BSE) signals were collected at an accelerating voltage of 20 kV under high vacuum. During BSE-EDS analysis, the detector deadtime was maintained at approximately 25%, and count rates exceeded 40,000 counts per second (cps). The collected data were analyzed using Aztec software (Version 3.0, Oxford Instruments, UK) to identify elements present in the samples.

2.2. Beneficiation Methods

Before the flotation experiments, 50 g of the sample was weighed and added to 1000 mL of tap water, corresponding to a pulp density of approximately 5 wt.% solids, then mixed at 1500 rpm for 15 min to remove any remaining BM. The relatively low pulp density was deliberately selected because, compared with conventional ores where higher pulp densities are common, this material contains a significant proportion of plastic particles and fine BM, which would otherwise increase slurry viscosity and promote unwanted entrainment at higher solids. Operating at lower pulp density results in a cleaner froth product and improves the selectivity of plastic flotation in the first stage.

The dispersed BM was separated using a 74 μm sieve, and the oversize material was fed into a 1.5 L self-aerated Denver-type flotation machine operating at an impeller speed of 1200 rpm. No pH modifiers were used to maintain a more environmentally friendly process. City tap water was utilized throughout the experiments, with the natural pulp pH measured at approximately 8.3.

The choice of collectors was guided by their well-documented performance in sulfide and base-metal flotation systems. KAX was included as a reference sulfhydryl collector to benchmark Cu flotation response. Aerophine-3418A, a phosphine-based collector, was selected for its known selectivity towards Cu minerals and lower affinity for iron and aluminum, which was expected to enhance Cu foil recovery with minimal contamination. Aerofloat-211 and Aerofloat-242 (dithiophosphates) were tested to evaluate whether stronger adsorption on Cu surfaces could further improve recovery or concentrate grade. Compared structurally to xanthates, but using phosphorus (P) instead of C, making a P-based sulfur (S) collector with improved selectivity. Sodium di-isobutyldithiophosphinate-based collector is primarily composed of sodium (Na), P, S, C, and hydrogen (H). Its approximate molecular formula is C₈H₁₈NaPS₂ [37,38]. Importantly, all collectors chosen are commercial reagents widely used in large-scale natural Cu ore processing, with well-understood chemistry and minimized environmental drawbacks, making the results more readily transferable to industrial LiB recycling operations.

Once the reagents had sufficient contact with the particles, the air valve was opened to introduce an airflow of 3 L/min, and the floated particles were collected during a 3 min flotation period. The first-stage flotation aimed to separate the naturally hydrophobic plastic fraction from the metallic components. MIBC was added in five equal charges of 100 g/t for each stage (5 mg per step for 50 g feed) 60 s before aeration, mixing at 1200 rpm (room temperature, ~22 ± 2 °C). Following plastic separation, Cu particles remaining in the sink fraction were floated using various collector reagents (KAX, 3418A, Aero-3739, Aerofloat-242, and Aerofloat-211). During Cu flotation, collectors were dosed totally at 3000 g/t (three equal charges of 1000 g/t for each stage, 50 mg per addition, 150 mg total per test) with 5 min conditioning at 1200 rpm after each addition. No additional cleaning or scavenger stages were applied to better assess the intrinsic selectivity of the flotation reagents under rougher-only conditions.

3. Results

3.1. Characterization Studies

Elemental analysis of the LiB sample used in the experimental studies revealed the presence of several key metals. As summarized in

Table 1. Metal contents and recoveries of the products after comminution and classification.

Products	Amount, wt.%	Cu		Al		Fe		Ni
		C 1, %	D 2, %	C, %	D, %	C, %	D, %	C, %	D, %
−2 + 0.2 mm	31.9	25.58	78.6	39.35	79.5	0.35	54.6	1.73	13.1
BM 1	12.3	2.34	2.8	2.80	2.2	0.10	6.0	3.06	8.9
BM 2	17.6	1.70	2.9	4.10	4.6	0.10	8.6	4.45	18.4
BM 3	19.0	4.81	8.8	4.67	5.6	0.20	18.6	5.87	26.4
BM 4	19.2	3.74	6.9	6.63	8.1	0.13	12.2	7.30	33.2
Total	100.0	10.38	100.0	15.78	100.0	0.20	100.0	4.23	100.0
Products	Amount, wt.%	Co		Li		Mn		Others
		C, %	D, %	C, %	D, %	C, %	D, %	C, %	D, %
−2 + 0.2 mm	31.9	6.26	7.4	1.54	18.8	1.56	12.4	20.96	19.8
BM 1	12.3	20.83	9.5	2.18	10.3	3.03	9.3	64.07	23.5
BM 2	17.6	36.30	23.6	3.54	23.7	4.02	17.6	44.37	23.1
BM 3	19.0	40.85	28.9	3.43	24.9	5.33	25.3	31.98	18.0
BM 4	19.2	42.90	30.6	3.05	22.3	7.37	35.4	27.40	15.6
Total	100.0	26.94	100.0	2.62	100.0	4.01	100.0	33.72	100.0

¹ Content, ² Distribution.

, the concentrations of Cu, Al, Co, Ni, and Mn in the initial feed material were 10.38%, 15.78%, 26.94%, 4.23%, and 4.01%, respectively. These metals originate primarily from current collectors (Cu, Al) and active electrode materials (Co, Ni, Mn). Following staged comminution and particle size classification, a significant redistribution of these metals was observed across different size fractions. Cu and Al, which serve as the primary current collector foils, became increasingly concentrated in the coarse fraction (−2 + 0.2 mm), representing approximately 25% of the feed mass. In this fraction, the combined Cu and Al content reached nearly 65 wt.%, demonstrating efficient liberation and size-based separation of the foil materials during mechanical processing.

By contrast, the fine fraction (−0.2 mm), referred to as “BM”, was significantly enriched in Co, Ni, and Mn, metals associated with cathode materials. The Co content increased steadily in finer BM fractions as size reduction progressed, rising from 20.83% in BM 1 to 42.90% in BM 4. This marked enrichment indicates progressive liberation and concentration of Co-bearing cathode materials during successive comminution and classification steps. This process likely reflects the effective detachment of Co from Al current collector foils, governed by the breakdown of strong polymeric binders (e.g., PVDF) at later milling stages.

The carbonaceous content (graphite), primarily coated onto Cu foil, was more readily detached during initial crushing steps due to its relatively weak interfacial bonding. This is evidenced by a high “others” fraction (mainly graphite, plastic, and residual binder) of up to 64.07 wt.% in BM 1. Moreover, it was observed that intensive comminution not only liberated the active materials but also fragmented the Cu and Al foils into fine particles, thereby contaminating the BM. This contamination was particularly pronounced in BM 3 and BM 4, where the intrusion of Cu and Al into the fine fractions increased significantly. These findings indicate that the multi-stage size reduction process causes foil-structured Cu and Al metals to migrate into the fine fractions, leading to contamination of the BM and potentially reducing overall metal recovery efficiency. A particle size distribution analysis was conducted on the oversize fraction, comprising mainly Cu, Al, plastic materials, and casing metals, retained on the sieve after comminution and classification. The particle size characteristics of this material, which was subsequently used in flotation experiments, were quantified by determining the d₈₀ and d₅₀ values, measured at 0.95 mm and 0.62 mm, respectively. These size parameters are critical for optimizing flotation performance, as they affect the hydrodynamic behavior and surface interactions of particles during the separation process.

As illustrated in Figure 3, the highest concentrations of Cu and Al were detected in the −0.8 + 0.5 mm size fraction. An increasing trend in the content of cathode active metals was observed toward the finer fractions. This suggests that finer Al foil particles retain a considerable amount of unreleased cathode material, indicating incomplete liberation during the comminution process.

SEM analysis revealed notable changes in surface morphology and composition of the foils after conditioning with 3418A. As shown in the SE images in Figure 4, the surface roughness of Cu foil increased remarkably after conditioning, suggesting the adsorption of 3418A collector. This could be supported by BSE-EDS results: as shown in Figure 5 and Figure 6, after conditioning, the surface C and S contents increased remarkably, from 7.6% to 26.3% and from non-detectable to 16.4%, respectively. Meanwhile, the Cu content decreased from 91.7% to 24.3%. Surface modifications on Cu foil are primarily due to the adsorption of di-isobutyldithiophosphinate species from 3418A, where the phosphorus–sulfur functional groups form strong chemisorbed complexes on the mineral surface, altering its flotation behavior. In comparison, no significant changes in the surface roughness and composition could be observed on Al foil after conditioning (as shown in Figure 4, Figure 5 and Figure 6), suggesting the collector did not adsorb onto its surface.

3.2. Flotation Experiments

3.2.1. Plastic Flotation

Plastics are naturally hydrophobic and possess a much lower specific gravity compared to metals. Fadel et al. [39] reported that pure PET exhibits a contact angle of 78.9°, confirming its hydrophobic nature. In froth flotation, selective separation is achieved based on differences in particle hydrophobicity. In our previous study on the selective recovery of metals from spent mobile phone LiBs [40], contact angle measurements were used to assess the surface hydrophobicity of key metallic constituents such as Cu and Al. The relatively low contact angles of unconditioned metallic foils, 50.5° for Cu and 56.1° for Al, support the effectiveness of plastic flotation, where a clear separation between plastics and metals can be realized. Plastic flotation tests were conducted in five stages: no frother was used in the first stage, followed by four stages with 100 g/t MIBC added each time. Without any frother, about 12% of the total plastics floated, increasing to approximately 79% flotation recovery with 200 g/t of MIBC. The cumulative weight of floated plastics systematically increased, reaching 95% at 400 g/t. Due to size reduction processes dominated by shear forces, metal and plastic components adopted plate-like morphologies and became largely liberated from one another. This morphological transformation provides a significant advantage during flotation, facilitating effective separation of plastic particles, which exhibit substantially higher hydrophobicity than metals, even at relatively coarse particle sizes. As shown in

Table 2. Metal content and recoveries of flotation.

Collector Type	Products	Amount wt.%	Cu		Al		Co		Ni		Mn		Others
			C 1, %	R 2, %	C, %	R, %	C, %	R, %	C, %	R, %	C, %	R, %	C, %	R, %
KAX	Plastics	14.3	1.25	0.7	12.70	4.7	4.58	10.4	1.30	10.7	1.28	11.8	74.29	48.8
	BM	8.5	1.24	0.4	6.30	1.4	20.18	27.3	5.66	27.6	5.14	28.1	50.96	19.9
	Cu	38.8	62.47	95.8	22.25	22.3	2.25	14.0	1.02	22.8	0.95	23.8	8.53	15.2
	Al	38.5	2.02	3.1	71.84	71.6	7.84	48.3	1.75	38.9	1.46	36.3	9.03	16.0
	Total	100.0	25.28	100.0	38.63	100.0	6.25	100.0	1.73	100.0	1.55	100.0	21.69	100.0
3418A	Plastics	15.7	1.41	0.9	12.68	5.1	4.96	12.5	1.58	13.9	1.46	14.5	73.79	54.0
	BM	10.0	1.33	0.5	6.28	1.6	19.96	32.0	5.51	30.7	5.41	34.2	51.81	24.1
	Cu	37.9	65.13	96.4	21.12	20.7	2.03	12.4	0.73	15.5	0.67	16.1	7.06	12.5
	Al	36.5	1.52	2.2	76.98	72.6	7.33	43.1	1.95	39.9	1.52	35.2	5.55	9.4
	Total	100.0	25.58	100.0	38.70	100.0	6.21	100.0	1.78	100.0	1.58	100.0	21.43	100.0
AERO 3739	Plastics	15.5	1.52	0.9	12.47	4.9	5.03	12.7	1.56	14.1	1.40	14.5	74.03	55.3
	BM	10.1	1.28	0.5	6.20	1.6	19.95	32.8	5.29	31.1	5.35	36.0	52.35	25.4
	Cu	5.9	63.27	15.0	21.97	3.2	1.83	1.8	0.74	2.5	0.63	2.5	8.69	2.5
	Al	68.5	30.25	83.6	52.25	90.3	4.71	52.7	1.31	52.3	1.03	47.1	5.11	16.9
	Total	100.0	24.81	100.0	39.66	100.0	6.13	100.0	1.72	100.0	1.50	100.0	20.77	100.0
AERO 242	Plastics	15.1	1.65	1.0	12.84	4.9	4.70	11.5	1.47	13.4	1.36	13.9	74.00	53.6
	BM	8.0	1.18	0.4	5.57	1.1	19.79	25.5	4.90	23.6	5.55	29.9	52.99	20.3
	Cu	21.4	65.85	56.6	15.93	8.5	1.56	5.4	0.94	12.2	0.78	11.3	12.33	12.7
	Al	55.5	18.88	42.0	61.60	85.5	6.43	57.6	1.52	50.8	1.20	44.9	5.06	13.5
	Total	100.0	24.94	100.0	39.96	100.0	6.19	100.0	1.66	100.0	1.48	100.0	20.86	100.0
AERO 211	Plastics	15.7	1.58	1.0	12.32	4.9	4.83	12.1	1.48	13.2	1.44	14.5	74.12	53.8
	BM	9.9	1.28	0.5	5.48	1.4	19.42	30.7	5.32	29.9	5.35	34.2	52.97	24.4
	Cu	28.6	69.22	80.1	16.21	11.8	1.82	8.3	0.68	11.0	0.60	11.1	8.82	11.7
	Al	45.9	9.91	18.4	69.79	81.8	6.68	48.9	1.76	45.9	1.36	40.2	4.75	10.1
	Total	100.0	24.70	100.0	39.11	100.0	6.26	100.0	1.76	100.0	1.55	100.0	21.55	100.0

¹ Content, ² Recovery.

, approximately 16% of the feed material to the first flotation stage was recovered as a plastic product. Approximately 5% by weight of fine-sized Al particles, which exhibit greater floatability than Cu, were transferred to the floated product, whereas Cu particles, with lower floatability than Al, accounted for around 1.5%. A small portion of cathode active materials, still attached to Al foils reported to the plastic fraction.

Particle surface wettability, whether hydrophilic or hydrophobic, can be effectively changed through surface chemical modifications [40]. Although Al particle surfaces are initially more hydrophobic than Cu surfaces before reagent addition, the hydrophobicity of Cu increases significantly (~95°) and exceeds that of Al (~78°) upon addition of collectors originally designed for natural Cu ore flotation [40]. In EoL LiBs, Cu and Al are typically present as thin, flat monolayer foils (see Figure 7). This unique morphology enables efficient separation through flotation techniques at significantly coarser particle sizes compared to those typically employed in the flotation of primary ores. Subsequently, the residual material within the cell, following plastic flotation, was processed through the Cu flotation circuit by introducing an appropriate dosage of collector and frother to facilitate the selective recovery of Cu-bearing components.

3.2.2. Copper Flotation

The sink fraction from the plastic flotation stage was subjected to Cu flotation using various collectors. 3418A exhibited the highest selectivity and recovery for Cu foils under the applied conditions. The Cu concentrate obtained with 3418A contained 65.13 ± 0.64 wt.% Cu with a recovery of 96.4 ± 0.10%, based on the average of three replicate chemical assays. These results confirm the high grade and recovery previously reported and demonstrate that the data are reproducible within a narrow confidence range.

The phosphorus–sulfur functional groups of 3418A are known to form strong Cu–S and Cu–P bonds, creating a hydrophobic layer that enhances bubble–particle attachment during flotation. The improved Cu floatability observed with 3418A can be attributed to its selective adsorption on Cu foil surfaces. SEM and BSE-EDS analyses (Figure 4, Figure 5 and Figure 6) revealed increased surface roughness and a notable rise in surface sulfur and carbon content after conditioning, suggesting adsorption of di-isobutyldithiophosphinate species. While these results provide qualitative evidence of surface modification, XPS or FTIR spectroscopy would be more suitable to confirm the chemisorbed collector species and clarify the binding mechanism at the molecular level. This finding is consistent with previous reports on the use of Aerophine collectors in complex sulfide ore systems [37,38]. In contrast, no notable adsorption or surface modification was detected on Al foils, confirming the selectivity of 3418A toward Cu. Furthermore, since the plastic components (PP/PE) are inherently hydrophobic with high contact angles (>75° [39]), they were floated in the first stage without collector addition, and any potential interaction with 3418A in the second stage was minimal. This selective adsorption behavior explains the high Cu grade and recovery obtained, as well as the efficient separation between Cu and Al fractions.

Another commercial reagent commonly preferred in the flotation of natural Cu ores, KAX, also yielded successful results, producing a Cu concentrate assaying 62.47% Cu with a 95.8% Cu recovery. Although the other three collectors produced concentrates with similar Cu grades, their recovery rates were lower due to the reduced quantity of floated material.

When evaluated based on Al content, the best performance was again achieved with the 3418A collector. A sink product containing approximately 77% Al was obtained with a 72.6% Al recovery. The relatively lower Al recovery compared to Cu is attributed to the loss of small-sized Al particles entrained in the Cu concentrate.

Graphite is not typically used in Al current collectors; however, it is predominantly employed as an anode material on Cu surfaces. Considering that some graphite remains adhered to Cu particles after comminution and screening, the “others” category for Cu and Al in Table 2 likely includes polymer-based separator fragments that were not effectively removed during the initial plastic flotation stage. With the addition of a collector in the Cu flotation circuit, a portion of these particles was recovered, whereas more fibrous and irregularly shaped separator fragments remained in the sink product, often attached to metallic particles. As a potential solution, these insulating plastic materials can be removed either by feeding the material into electrostatic or eddy current separators after drying or by subjecting them to a scavenger flotation circuit in the presence of suitable collectors and frothers.

3.3. Mass Balance and Overall Flowsheet

Based on the characterization and enrichment results, a comprehensive flowsheet was developed for recovering valuable metals from EoL LiBs sourced from EVs (Figure 8). Following an initial discharging stage, the EoL LiB sample underwent a four-stage comminution process. The BM, constituting 68.1% of the total weight, was then separated via sieving. The −2 + 0.2 mm particle size fraction was subjected to froth flotation, during which plastic components were selectively recovered using 500 g/t MIBC as a frother. The floated plastic fraction accounted for 5.0% of the feed and contained negligible metal content, making it suitable for disposal or potential secondary use in other industrial applications (e.g., energy recovery, composite filler production), thus reducing landfill burden.

The non-floated fraction was further processed in a Cu flotation circuit using 3000 g/t 3418A as the collector. This step produced a Cu concentrate representing 12.1% of the feed mass, assaying 65.13% Cu with 96.4% recovery, which constitutes a high-value feedstock for downstream hydrometallurgical or pyrometallurgical refining and supports efficient circular use of critical metals. The remaining tailings, constituting 11.6% of the initial feed and containing 76.98% Al, represent a secondary aluminum source that can be further upgraded by electrostatic or eddy current separation, potentially reducing reliance on primary aluminum production and its associated energy consumption.

The observed decrease in overall Cu recovery is primarily attributed to the removal of fine BM during the crushing and classification stages, which inevitably contains a fraction of fine Cu and Al current collector particles. To minimize such losses and improve the overall process efficiency, future work could focus on implementing additional pretreatment steps aimed at selectively scraping off the active electrode materials before size reduction. Potential approaches include mild thermal treatment to weaken the adhesion of the electrode coatings or controlled mechanical abrasion techniques to remove the BM without causing excessive fragmentation of the metallic foils.

To enhance Cu content and Al recovery, it is also recommended to reduce the floatability of Al particles through Sodium Hydroxide (NaOH) pre-treatment, followed by feeding them into a cleaning flotation circuit. Alternatively, reshaping the particles into more spherical forms via appropriate grinding techniques may facilitate the separation of Al and Cu using electrostatic or eddy current separation methods. Moreover, the Al-rich sink product obtained after Cu flotation, which remains unfloated due to its low hydrophobicity, can be further processed through magnetic separation to remove associated ferromagnetic metals. This step not only increases the purity of the Al fraction but also enables the production of a marketable magnetic product enriched in Fe and Ni.

Compared with conventional separation methods, the proposed two-step flotation process offers several advantages. Mechanical screening and density-based classification are generally less effective at liberating and separating plastics from metallic foils, often resulting in residual plastic content exceeding 15 wt.% in the metallic fraction [40]. In contrast, our process achieved more than 95% plastic removal, producing a clean Cu–Al feed for subsequent processing. It should be noted that the collector dosage applied in this study (3 × 1000 g/t) is higher than typical values used in conventional mineral flotation. This is attributed to the presence of binder residues, carbon coatings, and passive surface layers on the Cu foils, which require more reagents to achieve effective surface hydrophobization.

The approach used in this study can be contrasted with that of Saneie et al. [32], who investigated froth flotation for the recovery of Cu and Al from 18650-type cylindrical LiBs using a 1 mm cut size. Under such conditions, a substantial fraction of current collector foils is lost with the BM, reducing overall metal recovery potential. In the present work, a finer cut size of 0.2 mm was selected specifically to minimize these losses. Moreover, a multi-stage crushing and classification stage was applied to remove approximately 95% of the residual BM adhering to the foils before flotation. This combination of finer classification and pre-cleaning not only improves the selectivity of the subsequent flotation stages but also makes the process more representative of industrial practice for mixed NMC and LCO batteries from electric vehicles, where maximizing metal recovery and minimizing contamination of the plastic fraction are crucial objectives.

On a per-ton basis, the reagent usage of the proposed flowsheet is as follows: the Cu stage collector (3418A) corresponds to ≈3.0 kg/t feed (3 × 1000 g/t), while the frother (MIBC) consumption at the selected operating point is ≈0.5 kg/t. Based on typical procurement prices for laboratory and industrial use (3418A: $10–30/kg [41]; MIBC: $3–10/kg [42]), the total reagent cost is estimated at approximately $45–105 per ton of feed, with MIBC contributing around $1.5–5/t. According to the corrected mass balance, the Cu concentrate accounts for 12.1 wt.% of the feed at 65.13 wt.% Cu, corresponding to ~0.121 t concentrate/t feed carrying ~0.079 t Cu. The Al-rich sink assays ~77 wt.% Al. By removing plastics and fines before chemical processing, the pre-concentration step is expected to reduce acid consumption and furnace duty, thereby lowering operating costs and greenhouse-gas emissions relative to direct treatment of mixed feed [43].

From a scale-up perspective, the reagent consumptions reported here are within a range that is technically feasible for pilot- and industrial-scale operations, and the flowsheet can be directly adapted to continuous flotation circuits with minimal modification. Future work should include a pilot-scale verification to validate reagent dosages and recovery efficiencies under steady-state conditions.

By implementing flotation as a pre-concentration step, the amount of inert material entering downstream processes can be significantly reduced, leading to improved leaching kinetics, lower reagent consumption, and reduced energy demand during thermal steps. These results indicate that flotation-based pre-treatment is both technically and economically advantageous compared to conventional approaches.

4. Conclusions

In this study, more than 95% of the plastic fraction was removed in the first flotation stage using MIBC as a frother, producing a low-metal discard stream. In the second stage, Cu foils were selectively floated with 3418A, yielding a Cu concentrate assaying 65.13% Cu with 96.4% recovery. The remaining tailings were enriched in Al (~77%), representing a valuable secondary feed material for further upgrading using electrostatic, eddy current, or magnetic separation methods. These results confirm that plastics, Cu, and Al can be efficiently liberated and separated through appropriate pre-treatment and flotation strategies.

Beyond demonstrating the technical feasibility of this approach, the study also highlights several process optimization opportunities. The partial loss of fine Al particles to the Cu concentrate suggests that pre-conditioning steps such as NaOH washing or particle reshaping could further enhance Al recovery and concentrate purity. Integration of scavenger flotation or post-flotation physical separation could reduce entrained plastics and residual active material, further improving overall process efficiency.

While these results are promising, some limitations remain. Additional replicate flotation tests and statistical analysis are needed to confirm the robustness of the reported recoveries. Future work will explore alkaline pre-treatment (e.g., NaOH washing) or mechanical/thermal surface-cleaning methods to further reduce electrode coating contamination, as well as the integration of this approach with downstream hydrometallurgical processes for complete metal recovery. Finally, pilot-scale experiments should be performed to evaluate process scalability and economic feasibility under realistic industrial conditions, ensuring that the proposed approach can be effectively implemented in LiBs recycling plants and contribute to a more sustainable circular economy for critical metals.