The CaO Enhanced Defluorination and Air-Jet Separation of Cathode-Active Material Coating for Direct Recycling Li-Ion Battery Electrodes

Abstract

With the rapid growth of the lithium-ion battery (LIBs) market, recycling and re-using end-of-life LIBs to reclaim the critical Li, Co, Ni, and Mn has become an urgent task. Presently, high temperature, strong acid, and alkali conditions are required to extract blended critical metals (CM) from the typical battery cathode. Hence, there is a need for more effective recycling processes for recycling blended Li, Co, Ni, and their direct regeneration for re-use in LIBs. The goal of the offered paper is the development of recycling technology for degraded battery cathode-active materials based on the thermal decomposition of polyvinylidene fluoride (PVDF) using calcination and air-jet stripping of active materials. The proposed air-jet erosion method of calcined cathode material stripping from Al foil allows for the flexible industry-applicable separation process, which is damage-free for both particles and substrate. The CaO calcination air-jet separation process and equipment can significantly improve the PVDF decomposition and the separation efficiency of the cathode materials. It is demonstrated that low-temperature CaO calcination at 350–450 °C associated with air-jet separation of active material is characterized by low environmental impact, high purity of the recycled material, and low cost as compared to pyro- and hydrometallurgical methods.

1. Introduction

Pyrometallurgical (PM) and hydrometallurgical (HM) technologies are commonly employed to recycle spent LIBs. PM recycling involves high-temperature smelting, which is a process that consumes a significant amount of energy. HM recycling is based on leaching with acids or bases to dissolve all electrode elements and compounds. The leaching technology consumes substantial quantities of chemicals and generates considerable wastewater. Although large-scale, economically viable PM technology allows the recycling of metals such as Li, Co, Ni, etc., the process costs are high, and it is not applicable for recycling of cathodes such as LiFePO₄ (LFP) and LiMn₂O₄ (LMO). Therefore, there is an urgent need to develop cost-effective, energy-efficient, environmentally friendly, and more sustainable recycling methods. Direct recycling (DR) of the electrode materials, based on cell disassembling, defluorination process, and active material-current collector separation, presents a promising approach to obtaining precursor for following relithiation. Compared to PM and HM recycling routes, DR offers significant advantages in energy consumption, safety, cost, flexibility, and critical materials recycling [1]. In general, the practical implementation of state-of-the-art DR technologies utilizing various relithitation routes has the following shortcomings:

(i)

Lack of flexible automatic LIB sorting and disassembling technologies allowing the processing of various types of EV LIB modules and cells.

(ii)

Absence of sustainable industry-approved automatic large-scale separation of active material coating and foil from LIB electrode.

(iii)

The relatively low purity of recycled materials diminishes the quality of re-synthesized active materials, resulting in the application of additional separation (flotation and other) processing steps.

(iv)

Application of relithiation technologies with low productivity and without control of the technology parameters.

The active material defluorination and current collector separation processes are critical for sustainability and for obtaining pure recycled active material at the industrial scale.

Usually, the cathode-active material consists of 90 wt. % metal oxides, 5–7 wt. % carbon black conductive agent, and 3–5 wt. % organic binder. It is deposited on the aluminum foil current collector with a thickness of about 80–100 μm (Figure 1a,b).

The organic binder, polyvinylidene fluoride (PVDF), is widely used to join powder cathode materials to aluminum foil and consolidate oxide particles due to its relatively high bonding strength and chemical stability. However, destroying the PVDF joints and separating the aluminum foil and cathode materials is difficult. Figure 1b,c demonstrate that mechanical stresses cannot accomplish stripping of calcined active material from the Al current collector. Therefore, optimization of PVDF thermal decomposition regimes is required to destroy the PVDF bonds. Despite the considerable progress of technology development in recent years, significant issues, such as the purity of the recycled material and environmental pollution with HF, remain to be addressed in the decomposition of PVDF and separating cathode-active materials from spent lithium-ion battery electrodes. Scaling up the dissolution of the binder PVDF in the cathode materials with an organic solvent is difficult and expensive [2].

The PVDF thermal defluorination process during the calcination of cathode material at about 500–600 °C is an effective method because of its simplicity, convenience, and high productivity, allowing the realization of large-scale production. Thermal dissociation of PVDF molecules during calcination effectively removes the PVDF binder from the interparticle boundaries and the particle-Al current collector interface (Figure 1). However, PVDF decomposition releases hydrogen fluoride (HF) gas, causing side effects such as air pollution and equipment corrosion. The HF tail gas pollution must be avoided by absorption or reaction with alkaline (such as CaO) or other compounds. However, from our knowledge, only some work is devoted to resolving these challenges [3]. The study [4] proposed using calcium oxide (CaO) as a low-temperature reaction compound to accomplish the defluorination of PVDF in spent cathodes. CaO can cause the reactions (Equation (1)) of PVDF decomposition while eliminating the release of fluoride because the CaF₂ synthesis prevents HF creation. The defluorination process destroys the interparticle bond, allowing efficient mechanical stripping of the particles from the aluminum current collector (Equation (1)).

[–[CH₂–CF₂]–]_n + CaO→C_nH_m + H₂ + CaF₂

(1)

C (carbon black) + H₂O = CO + H₂; CO + CaO = CaCO₃

(2)

Additionally, carbon black dissociation is available according to reactions (Equation (2)). This process allows the purification of the NMC particles joined with carbon black. However, the effective temperatures of reactions (Equation (2)) are more than 900 °C in a Nitrogen protective atmosphere. Therefore, the carbon dissociation process may be performed after the Al current collector is taken away.

The recent state-of-the-art calcination/pyrolysis studies mainly focus on the PVDF decomposition kinetics and reaction mechanisms to recycle valuable metals and compounds [5,6]. However, the influence of the PVDF decomposition on the adhesion and cohesion strength of the cathode-active powder material layer has not been examined yet. Moreover, the lack of real active material calcination examination results hinders the development of large-scale, low-cost, low-energy consumption industrial technology. The environmentally friendly strategies of defluorinating the cathode materials on the aluminum foils have significant practical implications for recycling spent LIBs.

That is why the first goal of the paper is to examine conventional heat treatment and air-jet-based separation technology and define the parameters affecting the PVDF decomposition and separation of the cathode material and the aluminum foil. The effect of CaO reaction medium, reaction temperature, and reaction time on the process kinetics and its comparison with conventional high-temperature heat treatment will be performed.

The stripping of the calcined cathode materials from aluminum current collector foil is an important step influencing the purity and quality of the recycled material. The authors in [7] developed resonant acoustic vibration technology (RAV) for stripping cathode-active material from spent Li-ion battery electrodes. They demonstrated that the efficiency of RAV techniques is up to 92%. The comparison with other stripping methods (magnetic stirring, sonication, and brushing) reveals that RAV stripping of heat-treated electrodes provides relatively high active material quality. Moreover, the recycled cathode-active materials retain their original polycrystalline particle structures. However, it should be noted that the RAV and other stripping methods have certain drawbacks, such as (i) the application of the manual cutting of strips into smaller components and (ii) contamination of the recycled powder with Al and carbon black, (iii) the necessity of strong separation forces or shear stresses for powder stripping and (iv) a relatively low degree of stripping efficiency (<95%). That is why the development of more effective stripping methods looks to be important. Among them is an air-jet erosion process similar to wind-induced erosion, which is more likely to occur with soils, clays, and other powder sediments [8]. The effectiveness of stripping the calcined cathode material by the air-jet erosion process depends on air velocity and jet geometry. The low-pressure cold spraying equipment [9] fits well for this purpose. However, to our knowledge, there is no data about the examination of the air-jet erosion method application for the stripping step. Therefore, the second goal of the paper is to examine air-jet erosion separation technology.

2. Materials and Methods

The flowchart for the PVDF defluorination process and stripping of the cathode material from the spent LIB aluminum current collector using air and CaO powder is shown in Figure 2. The spent LIB cathode was extracted from the discharged, NaCl-deactivated, and disassembled battery cell. The dried spent LIB cells were manually cut to separate a plastic shell, cathode and anode electrodes, and separator. The cathode electrode was cut into a 20 mm × 100 mm ribbon and placed into a stainless-steel boat with CaO powder. The mass ratios of CaO/electrode materials were set as 8:1. The boat was pushed into the tube furnace and calcined for different times (0, 10, 20, and 30 min) at various reaction temperatures in the range of 50–600 °C in an air atmosphere at a heating rate of 10 °C/min and free cooled (to room temperature). After removing the cathode electrode from the CaO powder, the surface of the electrode was washed with deionized water to remove residual CaO powder. The calcined active material was manually stripped from the cathode current collector by a brush. The spent cathodes, calcined samples, and the obtained cathode material powder were weighed to calculate the calcination and separation efficiency of the proposed technology steps. Two types of cathode-active materials were examined: LiCoO₂ and LiNi_xCo_yMn_zO₂ (NCM).

The weight loss during heating is defined as the ratio m/m₀. The degree of conversion during calcination of cathode materials on the aluminum foil can be expressed as $\alpha ,$ and $\alpha$ was calculated with Equation (3).

$$\alpha ={\displaystyle \frac{m_0-m}{m_0-m_f}},$$

(3)

where m₀, m, and m_f are the initial, instantaneous, and final masses of the active material, respectively. The separation efficiency, s, was determined as the ratio s = m_s/m_f where m_s is the active material mass after stripping from the Al current collector.

TGA experiments were conducted using a Discovery TGA 5500 series instrument (TA Instruments, New Castle, DE, USA). Approximately 2 mg of each sample was loaded into individual pre-weighed TGA aluminum pans (TA Instruments, New Castle, DE, USA) before being placed onto the instrument’s sample platform. Subsequently, each sample underwent a temperature ramp from room temperature to 600 °C at a rate of 10 °C/min, with weight monitoring. The weight loss (%) for each sample was then plotted against the temperature of the furnace to generate its respective TGA curve. Additionally, the first derivative of the sample weight concerning the furnace temperature was computed to derive the first derivative curve for each TGA plot. Due to minimal fluctuations in sample weight during the experiment, the corresponding first derivative curves appeared highly noisy. To mitigate this noise, the curves were smoothed by averaging 150 adjacent points.

Characterizing active material adhesion strength to the Al current collector and interparticle bonding due to the binder at the interface is very important in evaluating defluorination effectiveness. From this viewpoint, applying classical adhesion strength measurement methods according to ASTM D905-08 [10] is irrelevant. The erosion resistance may be a suitable criterion of the calcined material’s mechanical properties because they are controlled by binder presence at the interface. The ASTM G76 test method may be used for these purposes. It utilizes an air erosion approach involving a small nozzle delivering a stream of gas that impacts the surface of a test specimen (Figure 3). The nozzle length-to-diameter ratio should be 25:1 or greater to achieve an acceptable particle velocity distribution in the stream. The nozzle (Figure 3a) consists of a tube about 1.5 mm inner diameter, l = 50 mm long, manufactured from an erosion-resistant material. The throat diameter is 1.5 mm, the S-nozzle-to-specimen distance (10 mm), and θ is the impingement angle (90 °C).

The test gas was dry air. The test temperature was ambient (between 18 °C and 28 °C). The air pressure was 6 bar (6 × 10⁵ Pa). Erosion resistance determination requires measurement of the eroded area on the specimen and is subject to microscope examination with a digital optical microscope VHX-2000 (Keyence, Osaka, Japan). The spot diameter and depth profile of an erosion crater produced were measured.

The surface morphologies and microstructures of the various materials were analyzed using MERLIN Compact (Carl Zeiss, Oberkochen, Germany) field emission scanning electron microscope.

3. Results

3.1. Defluorination Examination

The main purpose of low-temperature calcination is to scavenge the fluorine-containing contaminants in the cathode-active materials to mitigate their negative effects. The main ones are (i) active material contamination and (ii) environmental pollution with fluoride. The PVDF defluorination examination was focused on the kinetic mechanisms and the development of advanced CaO-assisted processing. The calcination results (m/m₀ = f(T)) for the NMC cathode are shown in Figure 4, together with the thermogravimetry (TG) curves of the electrode materials at the heating rate of 10 °C/min. There are two weight loss stages during the calcination process:

• The slight weight loss of 1–1.5% at the temperature range of 50 to 230 °C is ascribed to the volatilization of the electrolyte and the adsorbed water (volatilization stage).
• The weight loss of about 6.0–8.0% at the temperature range of 250 and 600 °C, which is ascribed to the decomposition of PVDF in the spent LIB cathode and the formation of volatile small molecular weight compounds such as vinylidene fluoride (decomposition stage) [11].

In general, experimental data agree with works [2,4,5,11]. However, the weight loss at the first low-temperature defluorination stage of NMC + CaO calcination is higher than that of NMC calcination. The main reason for such calcination behavior seems to be the volatilization of the organic electrolyte (vinyl carbonate, etc.) and the decomposition of the electrolyte lithium hexafluorophosphate (LiPF6) [12], which is accelerated due to harmless CaF₂ precipitate formation. The 120–130 °C volatilization of LIB electrolyte examination [12] reveals that 97.18% of F was fixed in the precipitates.

A comparison of the spent NMC cathode calcination data and thermogravimetric curves of the spent NMC active material powder calcination (Figure 4) demonstrates the following:

The small difference in the weight loss curves of NMC cathode coating and NMC powder stripped from the cathode reveals similar reaction processes for both cases.

The effect of CaO at the low-temperature calcination stage (electrolyte evaporation) of NMC cathode coating is higher than that of NMC powder.

The weight loss curves of NMC cathode coating at the high-temperature stage differ considerably for NMC and NMC + CaO cases, while the TGA curves for these cases are similar. This means that CaO diffusion processes through the NMC layer of 80 µm thickness play a significant role during calcination.

The data in Figure 5 demonstrate that CaO enhances the electrolyte defluorination reactions at the low-temperature stage. The weight loss increased from 1.14% to 1.60%. The influence of H₂O evaporation is negligible (see Figure 5c TGA curve for calcined NMC). The electrolyte and PVDF thermal decomposition reactions are complex [12], as illustrated by weight loss–temperature derivation curves (Figure 5). some peaks and valleys on these curves reveal this issue.

There are several approaches to modeling the thermal decomposition process during calcination. The simplest is the empirical model, based on global kinetics described by the Arrhenius equation. It is employed to evaluate the correlation of mass conversion rate with the calcination temperature. In this case, a general expression of the decomposition of solid material is

$${\displaystyle \frac{d\alpha }{dt}}=k\left(T\right)f\left(\alpha \right)$$

(4)

where f(α) is the function of α, where α, the degree of conversion at time t, is defined by Equation (3). f(α) is the function depending on the reaction mechanism. k(T) is the rate constant at the temperature T, which complies with the Arrhenius equation

$$k\left(T\right)=Aexp\left({\displaystyle \frac{-E_{\alpha }}{RT}}\right)$$

(5)

where A is the pre-exponential factor, E_α is the activation energy, T is the absolute temperature, and R is the gas constant. Temperature is related to time, t, as T_t = T₀ + βt where β is the heating rate. That is why the general equation to calculate the kinetic parameters is the following:

$${\displaystyle \frac{d\alpha }{dt}}=f\left(\alpha \right)Aexp\left({\displaystyle \frac{-E_{\alpha }}{RT}}\right).$$

(6)

The TGA data for LIB-active material defluorination stages shown in Figure 6 demonstrate that CaO greatly influences thermal destruction reactions. The conversion rate of electrolyte decomposition with CaO at 70 °C is higher by order than that of pure electrolyte. The NMC + CaO conversion rate at a temperature of 170 °C is approximately four times higher than that of NMC. It is necessary to note that there is no essential effect of CaO on the conversion rate in the temperature ranges of 70–200 °C for the first stage and 270–550 °C for the second stage. A detailed examination of the activation energy is required to define the reasons for such a thermal decomposition behavior, which will be discussed in future work.

The defluorination effect of spent LIB cathode calcination and the microstructure of calcined NMC powder are examined by SEM and EDS. The results presented in Figure 7 and Figure 8 demonstrate the structure and composition of spent NMC and LiCo₂ active materials, and Figure 9 and Figure 10 show the structure and composition of the NMC and LiCo₂ active materials calcined at 400 °C for 30 min.

The results shown in Figure 7a,b reveal the presence of small NMC crystals (approximately 20%) in the spent powder material stripped from the Al current collector. The polycrystalline spheres of 20–50 μm size contain various defects. Examination of the NMC chemical composition of crystals and polycrystals shown in table (Figure 7c) talks about the relatively uniform distribution of PVDF on the particle surface.

The SEM micrographs of spent LiCoO₂ active material (Figure 8a,b) demonstrate LiCoO₂ crystals of 5–10 μm size with some defects such as cracks and failure. The PVDF distribution in the powder matrix is heterogeneous (table of Figure 8c). Some areas of the crystal surface are not covered with PVDF.

The effect of NMC electrode calcination at 400 °C on the active material structure is shown in Figure 9a,b. It is seen that (i) small NMC crystals are agglomerated on the NMC spherical polycrystal surfaces, and (ii) there is no F content on the NMC crystal surface (table of Figure 9d). Therefore, the PVDF defluorination process is believed to be completed. Calcination with CaO does not influence the NMC structure (Figure 9c).

Calcination of the spent LiCoO₂ active material leads to similar effects (Figure 10a,b). Some small LiCoO₂ crystals are collected on the surface of big crystals. However, the small crystal aggregates occupy only small fields of the big crystal surface because the content of small crystals is substantially less than that of NMC crystals. It is important to note that PVDF decomposition in LiCoO₂ active material is incomplete. Some fluor-containing compounds are in the pores (table of Figure 10c).

3.2. Air-Jet Separation Technology Examination

Although many ways are available to recycle spent LIBs, the separation of the electrode materials from the current collector, either directly or indirectly, needs to be realized. These processes aim to overcome the adhesion between the electrode materials and the current collector (copper/aluminum foils) and cohesion interparticle bonding. The subsequent or simultaneous stripping of the active material from current collectors by mechanical processing such as sonication, stirring, curling–uncurling method and impact liberation are used [13,14]. The air-jet erosion method of stripping allows for careful control of the process of particle separation due to vortex generation at the air-jet-surface interface (Figure 3b). The surface topography of the erosion spot of LiCoO₂ active material (Figure 11a) calcined at 400 °C for 30 min demonstrates that the erosion depth varies because PVDF binder films have not entirely decomposed, as stated above. An increase of LiCoO₂ calcination temperature to 450 °C results in the elimination of the whole material from the erosion spot area (Figure 11b).

4. Discussion

4.1. Defluorination

The experimental results described in p.3.1 demonstrate that the low-temperature calcination method patented in 2003 [15] may be an effective technology for separating organic materials. Further development of the Li-ion battery electrode recycling process by association of calcination and hydrometallurgical techniques results in the combined recovery of cobalt and lithium from spent lithium-ion batteries [16,17,18,19]. The results of vacuum pyrolysis of cathode material showed that the cathode powder of LiCoO₂ and CoO was separated from aluminum foils at the following regime: temperature of 600 °C, vacuum evaporation time of 30 min, and residual gas pressure of 1.0 kPa [16]. However, the vacuum pyrolysis at a temperature of 600 °C leads to contamination of the recycled cathode black mass with Al because the temperature is near the melting of Al.

The effectiveness of the thermal decomposition of the polyvinylidene fluoride binder depends on the diffusion processes and detachment of the active material coating from the current collector during the loading of the remaining particulate agglomerates [17,19]. The authors [17] compared the calcination-separation process with the mechanical recycling process that utilizes a cutting mill to separate the current collector and coating. They found that only 97.1% w/w of the electrode coating was obtained with aluminum impurities of 0.1% w/w, 30 times purer than the comparative process. The contamination level remains relatively high, and diminishing the defluorination reaction temperature is one of the main ways of improving the calcination technology.

The two-stage kinetics of the defluorination process of Li-ion battery electrodes (Figure 4 and Figure 6) defined by TGA [4,5,11,20] allows the careful controlling of the low-temperature calcination process as the more advantageous industrial production technique than organic leaching because of its low price and simple operation. However, determining the kinetic parameters for each case is not easy and is based on the traditional models [5,11,21]. For example, the interaction between fluorine-containing substances and LIB-active materials studied by Hang et al. [22] results in active porous materials adsorbing the residues of electrolyte on the particle surface and into the pores (at 20–200 °C), while PVDF forms a liquid film and covers the AM particles surface at the temperatures of 400–500 °C. These reactions hinder the effective removal of PVDF, leaving fluorine-containing contaminants in active materials. The authors in [22] found that the possible fluorine contamination of the calcined black mass varies from 1.4 wt. % to 3.7 wt. % and depends on the calcination temperature. Moreover, the fluorine-containing contaminants cannot be removed completely by increasing the calcination temperature. We defined the similar effect of the presence of fluoride compounds in the pores (≈0.89 wt. %) of the LiCoO₂ active material calcined at 400 °C (Figure 10c). A detailed examination of this effect and other options of defluorination technologies similar to those described in [23,24] will be made in future work.

The continuous purification of the calcination atmosphere is a very important feature of the defluorination technology step. Reactions (1) and (2) eliminate hazardous gas (hydrogen fluoride) from the calcination atmosphere. The calcium hydroxide solution removes hydrogen fluoride from the pyrolitic atmosphere [24]. The reaction (2) forms the CaF₂ compound, which may be re-used in various fields [25,26,27,28]. Applying CaO-CaF₂ powder mixtures for glass-ceramics composites may be a good option for re-using this product of the defluorination reaction (2). The authors in [26] showed that adding CaF₂ enhanced the glass crystallization and increased the mechanical properties of composites. The content of CaF₂ in these composites is about 7 wt. %. Therefore, the application of the products of the defluorination reaction (2) looks to be promising.

4.2. Air-Jet Separation Technology

The effect of CaO-assisted calcination of NMC active material on the air-jet erosion spot geometry is seen in the graph of Figure 12. CaO accelerates the PVDF decomposition reactions at temperatures 100–400 °C. However, this effect is not seen for calcination at 450 °C. The air-jet erosion spot geometry depends on the kinetics of PVDF defluorination, which is the function of temperature and CaO content. That is why SEM micrographs of the LiCoO₂ cathode cross sections (Figure 13) illustrate that air-jet assisted erosion allows easy stripping of calcined cathode material from Al substrate without any foil defects (Figure 13b). Some particles are joined with the substrate because of not full decomposition of the PVDF binder at the calcination temperature of 400 °C for 30 min. However, the obtained results demonstrate the potential of air-jet erosion separation of active material, considering that the air-jet erosion parameters have not been optimized yet. The additional important result is that SEM micrographs of the separated material and substrate do not show any damage resulting from the erosion step. Therefore, it will lead to low contamination of the recycled active material with the substrate metal.

Two apparent factors define the air-jet erosion process: (i) bonding at the particle-particle and particle-substrate interface and parameters of the air-jet (velocity and temperature). Van der Waals force plays the main role in the bonding [14]. The chemical bond between PVDF and Al foil needs to be broken through the decomposition of PVDF, which depends on the kinetics of defluorination. The process simulation and theoretical calculation results [14] demonstrate that the interaction between PVDF and active material particles/substrate depends on active material powder composition. The results mentioned above showed significant differences in the distribution, binding force, and action mechanism of PVDF for different cathode materials, implying that the separation efficiency for different LIBs would be considerably different even with the same separation technology. Therefore, the task of ongoing work is examining the air-jet erosion separation process for various cathode-active materials.

4.3. Defluorination Reactor

Based on the discussed results, the continuous calcination-separation route and equipment (Figure 14) will be designed and tried in the pilot regime.

The main features of the proposed two-chamber reactor are the following:

○

horizontal metal wire mesh conveyors, which are a cost-effective solution for efficient transport of the spent LIB electrodes,

○

two air-jet De-Laval guns for stripping of calcined active material,

○

cyclone collector for collecting the processed active material and

○

filtering unit.

5. Conclusions

The organic binder PVDF in the LIBs provides the proper cohesion and adhesion of cathode-active materials and their electronic contact with current collectors. However, the existence of fluorine-containing compounds results in difficulties in the separation of cathode material and Al foil and environmental pollution during the recycling process for spent LIBs. The results demonstrate that the calcination of NMC and LiCoO₂ cathodes with CaO leads to effective PVDF decomposition at 400–450 °C, preventing environmental pollution due to the formation of the harmless CaF₂.

The decomposition process consists of two stages: electrolyte volatilization and PVDF defluorination. Their kinetics dramatically depend on the temperature distribution in the active material layer and the diffusion of CaO into this layer. The TGA examination demonstrates the dependence of the conversion rate on the temperature and presence of CaO.

The proposed air-jet erosion method for calcined cathode material stripping from Al foil allows for a flexible separation process that is damage-free for both particles and substrate.

The proposed CaO calcination air-jet separation process and equipment can significantly improve the PVDF decomposition and separation efficiency of the cathode materials.