The rapid proliferation of electric vehicles and portable electronics has precipitated a corresponding surge in the global production and, consequently, the eventual disposal of lithium-ion batteries. The spent Li-ion battery represents a significant secondary resource for valuable metals like lithium, cobalt, nickel, and copper, while also posing substantial environmental hazards if not managed properly. Among various recycling methodologies—pyrometallurgical, hydrometallurgical, and mechanical-physical—the mechanical route is often favored for its lower energy consumption and reduced secondary pollution. A critical step within this mechanical recycling chain is the separation of different material fractions from the shredded battery black mass. Eddy current separation (ECS) has emerged as a pivotal technology for this purpose, leveraging the differences in electrical conductivity between metallic (e.g., aluminum, copper) and non-metallic or less conductive materials (e.g., lithium iron phosphate, carbon).
Despite its apparent simplicity, the industrial application of eddy current separation for complex streams like spent Li-ion battery fragments often yields suboptimal results. This inefficiency stems from the multitude of interacting operational and design parameters whose effects are not fully quantified. Parameters such as particle size, shape, feed rate, and magnetic roller configuration collectively influence the complex electromagnetic forces governing particle trajectory. This article delves into a detailed investigation of these factors, employing numerical simulation and theoretical modeling to elucidate their impact on separation efficiency for crushed lithium iron phosphate Li-ion battery material. The goal is to establish a robust framework for optimizing ECS performance in the critical context of Li-ion battery recycling.
Fundamentals of Eddy Current Separation
Eddy current separation is a non-contact technique that exploits the principles of electromagnetic induction. The core component is a rapidly rotating magnetic roller, typically constructed with an array of permanent magnets (e.g., NdFeB) arranged with alternating poles around a ferrous core. This configuration generates a time-varying (alternating) magnetic field in the space around the roller.
When a particle with significant electrical conductivity (like aluminum or copper from a spent Li-ion battery) enters this dynamic magnetic field, eddy currents are induced within the particle according to Faraday’s law of induction. These eddy currents, in turn, generate a secondary magnetic field whose polarity opposes the primary field from the roller (Lenz’s law). The interaction between the primary and secondary magnetic fields produces a repulsive force, often termed the eddy current force or magnetic drag force. For a particle of mass \(m\) and volume \(V\), the general expression for the time-averaged eddy current force \(\mathbf{F}_e\) in the direction normal to the roller surface can be expressed as:
$$
\mathbf{F}_e \propto \sigma V B^2 v_{rel}
$$
where \(\sigma\) is the electrical conductivity of the particle, \(B\) is the magnetic flux density, and \(v_{rel}\) is the relative velocity between the particle and the magnetic field. Non-conductive or poorly conductive particles (e.g., plastics, electrode active materials like LiFePO₄) experience negligible force and follow a ballistic trajectory dictated mainly by gravity and initial velocity. Conversely, conductive particles are deflected from this standard trajectory, enabling their physical separation. The success of this separation hinges on whether the vertical component of the eddy current force exceeds the gravitational force at a point before the particle leaves the roller’s influence.
Methodology: Simulation and Model Development
To systematically analyze the ECS process for spent Li-ion battery materials, a combined approach of computational simulation and analytical modeling was adopted. The investigation focused on aluminum particles, a key conductive component from the battery casing and current collectors.
Static Magnetic Field Simulation
The spatial distribution of the magnetic field generated by the roller is the foundational element determining the eddy current force. A 3D model of the magnetic roller assembly was constructed in COMSOL Multiphysics® software using the AC/DC Module. The roller consisted of NdFeB35 permanent magnets mounted on a steel core. Key geometric parameters included the core radius \(R_c\), the magnet (roller) radius \(R\), the number of magnetic pole pairs \(k\), and the roller length \(L\). The surrounding air domain was sufficiently large to avoid boundary effects. The simulation solved for the magnetostatic field to map the magnetic flux density \(\mathbf{B}\) in the region of interest. This static analysis was crucial for identifying design parameters affecting field strength before introducing motion.
Dynamic Force and Motion Modeling
To understand particle dynamics, analytical models for the eddy current force were derived for two canonical particle shapes relevant to shredded battery waste: a spherical particle and a thin circular disk (representing flake-like fragments).
For a spherical conductive particle of radius \(a\) entering the alternating magnetic field, the induced eddy current distribution is complex. Under assumptions of a uniform external field gradient and neglecting demagnetization effects, the resultant repulsive force \(F_{sph}\) at a point defined by the detachment angle \(\alpha\) (angle from the vertical) can be modeled as:
$$
F_{sph}(\alpha) = \frac{B_r^2 k (\omega_m R – v) a^7}{12 \rho R^3} \cdot \frac{1}{[\sec(\alpha – 1)]^2}
$$
where \(B_r\) is the radial component of the magnetic flux density, \(\omega_m\) is the angular velocity of the magnetic roller, \(v\) is the particle feed velocity, \(\rho\) is the resistivity of the particle material, and \(R\) is the roller radius.
For a thin circular disk of radius \(R_0\) and thickness \(H\), with its face oriented perpendicular to the field, the effective area for flux change is larger. The force \(F_{disk}\) can be expressed as:
$$
F_{disk}(\alpha) = \frac{B_r k (\omega_m R – v) S^2 H R_0^2 B_m}{16 \pi^2 \rho R^3} \cdot \frac{1}{[\sec(\alpha – 1)]^2}
$$
where \(S\) is the cross-sectional area of the disk and \(B_m\) is the internal magnetic field intensity within the particle.
The motion of the particle on the conveyor belt leading to the roller is governed by a balance of forces: the eddy current force \(\mathbf{F}_e\), gravity \(m\mathbf{g}\), the normal force from the belt \(\mathbf{N}\), and kinetic friction \(\mathbf{F}_f = \mu_k \mathbf{N}\), where \(\mu_k\) is the coefficient of kinetic friction. Crucially, the horizontal component of the eddy current force opposes the friction force. A critical “equilibrium angle” \(\beta\) exists where these horizontal forces balance:
$$
F_e(\beta)\sin\beta = \mu_k (mg – F_e(\beta)\cos\beta)
$$
For \(\alpha < \beta\), friction dominates and the particle’s horizontal velocity equals the belt speed \(v\). For \(\alpha > \beta\), the horizontal eddy force component exceeds friction, causing the particle to decelerate before finally detaching at the “detachment angle” \(\alpha_d\), where the vertical component of \(F_e\) equals the gravitational force \(mg\cos\alpha_d\). The trajectory post-detachment is parabolic, determining the separation bin.
Analysis of Influencing Parameters
The developed models allow for a systematic investigation of how various parameters affect the detachment angle \(\alpha_d\) and, consequently, the separation distance. The effects are summarized in the table below and discussed in detail.
| Parameter Category | Specific Parameter | Effect on Eddy Current Force / Detachment Angle | Physical Reason |
|---|---|---|---|
| Magnetic Roller Design | Number of Pole Pairs (\(k\)) | Negative Correlation | More poles decrease the spatial wavelength of the field alternation, reducing the effective field gradient and flux change rate for a given particle size. |
| Roller Radius (\(R\)) | Positive Correlation | A larger radius increases the region of high field intensity and provides a longer acceleration path for the particle. | |
| Roller Length (\(L\)) | Negligible Effect | The magnetic field in the region of interest is primarily determined by the 2D pole cross-section; length does not affect field strength perpendicular to the axis. | |
| Operational Conditions | Roller Rotational Speed (\(\omega_m\)) | Positive Correlation | Higher speed increases the rate of magnetic flux change (\(dB/dt\)), directly amplifying the induced eddy currents and the repulsive force. |
| Particle Feed Velocity (\(v\)) | Negative Correlation | A higher feed velocity reduces the relative velocity (\(v_{rel} = \omega_m R – v\)) between the particle and the magnetic field, thereby decreasing the induced force. | |
| Particle Characteristics | Particle Size (e.g., \(a\), \(R_0\)) | Strong Positive Correlation | Larger particles have greater volume for eddy current generation and experience a stronger force due to the \(a^7\) (sphere) or \(R_0^2\) (disk) dependency. |
| Particle Shape (Sphere vs. Disk) | Dominant Influence | Thin, flat disks present a larger effective cross-sectional area to the changing flux compared to spheres of equivalent mass, resulting in significantly higher forces. | |
| Electrical Conductivity (\(\sigma\)) | Positive Correlation | Higher conductivity (lower \(\rho\)) leads to stronger induced eddy currents for a given changing magnetic flux. | |
| System Interaction | Conveyor Belt Friction (\(\mu_k\)) | Negative Correlation | Higher friction increases the equilibrium angle \(\beta\), causing earlier deceleration and a lower velocity at detachment, which reduces the effective \(v_{rel}\) and the eddy current force. |
Detailed Discussion of Key Factors
1. Particle Size and Shape: These are the most significant material-related factors. The models reveal a power-law relationship between force and size, making separation efficiency highly sensitive to the size distribution of the shredded spent Li-ion battery feed. Pre-classification (e.g., screening) is often essential. Shape is even more critical; flaky aluminum from foil current collectors will be ejected much more effectively than spherical or lumpy metallic particles of the same mass. This necessitates an understanding of the liberation and morphology of materials during the initial shredding of the Li-ion battery.
2. Magnetic Roller Parameters: The simulation confirmed that while the number of pole pairs increases the frequency of field alternation, it dilutes the peak field strength available for induction. An optimal \(k\) exists for a given target particle size. The roller radius \(R\) positively influences the force, but increasing it has practical limits regarding cost and machine size.
3. The Critical Role of Friction (\(\mu_k\)): Traditional models often overlook the impact of the conveyor belt’s kinetic friction. Our analysis incorporates it through the equilibrium angle \(\beta\). A lower \(\mu_k\) is beneficial as it allows the particle to maintain a higher horizontal velocity closer to the detachment point, maximizing \(v_{rel}\). This can be achieved by using belts with low-friction coatings or by ensuring the feed material is dry. The relationship can be visualized by solving the equilibrium equation for different \(\mu_k\) values:
$$
\beta = \arctan\left(\frac{\mu_k mg}{F_e(\beta) + \mu_k mg \cot \beta}\right)
$$
An iterative solution shows that \(\beta\) decreases with decreasing \(\mu_k\), leading to less deceleration and a more forceful ejection. This insight is crucial for optimizing real-world ECS operations for spent Li-ion battery recycling.
Implications for Spent Lithium-Ion Battery Recycling
The findings from this modeling study have direct implications for optimizing the recycling chain for spent Li-ion batteries. The strong dependence on particle size and shape underscores the importance of the upstream pre-processing stages—specifically, the mechanical crushing and sieving steps. To achieve high-purity concentrates of aluminum and copper via ECS, the liberation of these metals from other battery components (e.g., active material coating, separator) must be effective, and the resulting fragments should ideally be flat and metallic-rich.
Furthermore, the operational parameters of the ECS unit must be tuned according to the characteristics of the feed material derived from the spent Li-ion battery. For instance, a feed consisting of finely shredded, fluffy material may require a different roller speed and feed rate configuration compared to a feed of coarser, denser granules. The explicit inclusion of belt friction in the model provides a previously underexplored lever for process improvement, suggesting that investments in high-quality, low-friction conveyor systems can yield tangible gains in separation efficiency and metal recovery rates.
Conclusion
Eddy current separation remains a vital technology within the mechanical recycling paradigm for recovering valuable metals from spent Li-ion batteries. However, its efficiency is governed by a complex interplay of design and operational parameters. This work has presented a consolidated analysis using COMSOL-based magnetic field simulation and derived analytical force models for spherical and disk-shaped particles. The study quantitatively highlights that particle size and shape are the predominant material factors, with thin, flat metallic fragments being far more susceptible to separation than compact ones. Key machine parameters like magnetic roller radius and rotational speed show a positive correlation with separation force, while the number of pole pairs and the particle feed velocity show a negative correlation.
A significant contribution of this analysis is the formal incorporation of conveyor belt kinetic friction into the particle motion model. It demonstrates that a lower friction coefficient favorably alters the force balance, allowing particles to experience a stronger eddy current repulsion at the point of detachment, thereby enhancing separation distance and purity. For recyclers processing end-of-life Li-ion batteries, these insights provide a clear guide for optimizing both the preparatory shredding stages to produce favorable particle morphologies and the operational settings of the ECS unit itself. Future work should focus on experimental validation with real shredded Li-ion battery fractions and extend the modeling to multi-particle interactions and non-ideal particle shapes to further bridge the gap between theory and industrial practice.