The expanding market for electric vehicles requires lithium-ion batteries (Li-ion batteries), as these are energy storage devices with high power, high capacity, high charging rates, and long life stability. The most common cathode materials for Li-ion batteries in electric vehicles [1
] belong to a group of layered mixed transition metal oxide compounds with rhombohedra symmetry (D3d5
space group), LiNix
]. Among them, the most popular and widely used material in Li-ion batteries is the phase LiNi0.33
Global demand for powerful, rechargeable Li-ion batteries, particularly for electric vehicles, has increased the demand for essential elements Co, Ni, Mn, and Li. For years, Co has been the most valuable and critical raw element needed in battery metals [3
]. In 2005, 25% of the end products made with cobalt in the EU were used in the manufacturing of battery chemicals, and this value had increased to 42% in 2015 [4
]. Therefore, end-of-life Li-ion batteries have become an essential secondary source to cover the Co requirements and for other necessary elements.
There are two basic technologies for the recycling of spent Li-ion batteries: The pyrometallurgical and the hydrometallurgical process [5
]. The pyrometallurgical route is a smelting process in which spent Li-ion batteries are entirely melted down without further pretreatment. Organic materials are needed as fuel to maintain the melting process. A molten metal phase consisting of Co, Cu, and Ni is obtained from the smelter and leached in acid solution after cooling. After separation and chemical precipitation, the metals are recovered as inorganic salts. Al, the substrate material of the cathode, Li, and Mn are oxidized, and they are combined with other metal oxides as a slag that can be treated to recover Li. Recently, a combined route consisting of mechanical treatment (comminution of the cathodes and mechanical separation of the Al substrate foils and NMC) and hydrometallurgical treatment (dissolution of the NMC and metal recovery therefrom) was described [14
]. The pyrometallurgical route is energy-intense due to the high temperatures required. It emits dust and hazardous gaseous compounds and results in a significant loss of materials. The hydrometallurgical route comprises the leaching of dismounted or shredded Li-ion batteries in strong inorganic acids to dissolve any metals and battery materials. The major challenge in this process is the separation of the ions from the concentrated metal ion liquor, mainly by selective precipitation, electrochemical deposition, or other techniques, such as solvent extraction or ion exchange. As in the pyrometallurgical route, the metals are obtained as inorganic salts. Losses typically occur from insufficient leaching and the precipitation/separation efficiencies of the metal salts.
Our motivation was to avoid the pyrometallurgical and hydrometallurgical recycling routes, as well as their specific drawbacks. In this paper, we described the basics of an alternative approach to recover NMC while preserving its chemical, physical, and morphological properties, with a minimal use of chemicals. This approach, which is designated as functional recycling (Figure 1
), can be applied to the cathodes of dismantled and separated end-of-life Li-ion batteries, as well as to residues or scrap from the production of cathodes. Based on the known design of a cathode in Li-ion batteries (Figure 1
), the first step in functional recycling is the complete removal of the cathode coating (which consists of NMC, a binder, and conductive carbon black) from the Al substrate. This requires arranging for the medium to be as minimally reactive as possible toward the NMC or the Al substrate foil, and to keep the contact time with the medium as short as possible, so that the NMC particles do not experience any degradation. The development of such a procedure was presented in this work. Basically, this treatment is followed by a second step, not considered here, in which the mechanically separated coating is dried and mechanically comminuted to release and separate the NMC particles from the binder/carbon black mixture.
2. Materials and Methods
NMC material and cathodes: The chemical composition of the NMC material used in this study was determined by an inductively coupled plasma-optical emission spectrometer (ICP-OES) to be Li0.945±0.007
, which was close to the ideal NMC stoichiometry. The NMC consisted of spherical agglomerates (c.f. Section 3.1
), designated as secondary particles, with diameters between 5 µm and 15 µm, which were constituted from primary particles with diameters of 0.5–1 µm (c.f. Section 3.1
). The studied cathodes consisted of 20 µm thick Al foils, and they were coated with a layer consisting of NMC, the organic binder polyvinylidene fluoride (PVDF), and carbon black as a conductive additive. Two different cathodes were studied with a mean single-sided coating layer thickness of 50 µm and 22 µm, and they were designated as cathode 1 and cathode 2, respectively. For cathode 1, the mass fraction of Al amounted to 14.1% of the total weight, and the total Li content was 56.84 mg per 1 g cathode. The mass fraction of Al in cathode 2 amounted to 27.4%, and the total Li content was 45.64 mg per 1 g cathode. A mechanically shredded fraction, as well as a preselected fraction of the cathodes with a mass fraction of 7.0% Al, were used.
Chemicals and reagents: All acids (HCl, HNO3) and chemicals (citric acid C6H8O7, KH2PO4, K2HPO4, NaOH, Na2CO3, and NaHCO3) used in this study were of analytical grade and were purchased from Merck (Darmstadt, Germany). Solutions of C6H8O7, KH2PO4, K2HPO4, and NaOH were prepared by dissolving the chemicals in deionized water (pH between 5.8 and 6.5, 18MΩ cm−1, Milli-Q, Darmstadt, Germany). The H2PO4−/HPO42− buffer solutions were prepared by mixing solutions of KH2PO4 (c = 66 mmol L−1) and K2HPO4 (c = 66 mmol L−1). Li, Ni, Mn, Co, Al, and S single-element standard solutions from Merck were used to prepare multielement calibration standards for ICP-OES measurements. The calibration standards for the measurements by ion chromatography (IC) were prepared from an SO42− standard solution from Merck.
Leaching experiments: The cathode foils were gently cut with scissors into square pieces with a size of 5 × 5 mm2. Pieces with a damaged coating layer or uneven edges were rejected. Weighed samples of the NMC (the cathode pieces or the shredded fraction) were put in contact with the leaching media. If not otherwise stated, the experiments were performed at 25 °C with a 50 mL volume of the leaching medium. All the leaching experiments were stirred at 400 min−1 using a magnetic stirrer (H+P Labortechnik AG, Oberschleißheim, Germany). During leaching, suspension samples were collected, and after filtration, the metal concentrations were analyzed by ICP-OES.
Sample digestion: The NMC starting material was dissolved in a mixture of HNO3, 69% (m/m), and HCl, 37% (m/m), using a high-pressure microwave digestion system (MLS Ethos Start, Leutkirch, Germany, 100 mL polytetrafluoroethylene (PTFE) vessels, maximum temperature 220 °C), and they were subsequently analyzed by ICP-OES.
Chemical analysis: An inductively coupled plasma-optical emission spectrometer (ICP-OES) with a dual-view option (iCap 6500 DUO, Thermo Scientific, Dreieich, Germany) was used to determine the elemental composition of the NMC and the concentrations of the dissolved elements. The sample introduction system was equipped with a parallel path nebulizer made of polyether ether ketone (PEEK) (MiraMist, Burgener Inc., Mississauga, ON, Canada), a cyclonic spray chamber (Glass Expansion, Port Melbourne, Victoria, Australia), and a ceramic injection tube (Glass Expansion, Port Melbourne, Victoria, Australia).
Ion analysis: Concentrations of sulfate ions were determined using ion chromatography (881 compact IC pro, Deutsche METROHM GmbH & Co. KG, Filderstadt, Germany). An METROSEP ASupp5 (Deutsche METROHM GmbH & Co. KG, Filderstadt, Germany; length 250 mm, diameter 4.0 mm) was used as a separation column, with a carbonate eluent consisting of 3.2 mmol L−1 Na2CO3 and 1.0 mmol L−1 NaHCO3. Samples and standard solutions were purified from metal ions using an SPE-H+-cartridge (MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany).
Raman confocal microscopy: NMC particles were studied with a confocal Raman microscope (DXR SmartRaman, Thermo Fisher Scientific, Dreieich, Germany) in the backscattering configuration. The microscope was equipped with a 532 nm excitation laser and a 900 grooves/mm grating to record the Raman spectra in the wavenumber range of 150–1250 cm−1. The incident laser light, generated with a laser power of 0.5 mW, was focused on the sample surface through a 100× microscope objective. The laser spot had a diameter of 1.6 µm.
SEM-EDX measurements: The materials were studied with SEM-EDX (scanning electron microscopy energy-dispersive X-ray spectroscopy, using a ZEISS EVO M 15 (Carl Zeiss Microscopy GmbH, Jena, Germany) equipped with EDAX TEAM™ EDS (AMETEK, Weiterstadt, Germany).
If NMC comes into contact with water, a protonation of the NMC surface takes place, which leads to M–OH2+ species at the surface. The enrichment of the positive charge at the surface requires a charge balance, which leads to the deintercalation of Li+. In practice, this refers to an exchange of Li+ vs. H+, which reduces the H+ ion concentration in solution and thus increases the pH value and leads to the dissolution of Al. However, not all the NMC is available for protonation due to the embedding in the PVDF polymer. Furthermore, diffusion processes, partial under-etching of the coating, and cracks and deformations can influence the protonation and deintercalation of Li, as well as the dissolution of Al. This means that the defined correlations between the solution contents of Li+ and Al3+, as well as the time-dependent leaching behavior, can hardly be reproduced to a high quality. If the Al is initially dissolved as a hydroxoalumina complex, the continued dissolution of the Al leads to a reduction in the hydroxide ion concentration, which finally leads to the precipitation of Al(OH)3, and possibly LiAl(OH)4. The insoluble Al(OH)3 precipitates in several or a few thick layers on the NMC particles. This passivation is considered to downgrade the electrochemical performance of the recovered NMC particles and might prevent their re-use.
The precipitation of Al(OH)3 on the NMC particles can be avoided if alkaline solutions with sufficient buffer capacity are used instead of water. Although an attack on the Al substrate film also occurs in the alkaline medium, the Al remains in solution in excess of the hydroxide ions, forming highly water-soluble hydroxoaluminate complexes. The alkaline medium leads to a partial attack on the NMC, which manifests itself in low solution concentrations of Ni, Mn, and Co. The compromise between the side reactions of dissolving Al and dissolving Ni, Mn, and Co is the choice of a pH value between 7 and 8. At the end of the experiments, the NMC particles are almost identical in size and morphology as the starting material used.
The amount of Li released in the experiments was small in relation to the total Li content in the studied cathode samples. Two parallel proceeding mechanisms seem to determine the release of Li: (i) The degradation of the NMC according to Billy et al. [15
] and (ii) a release via the dissolution of traces of inorganic salts, in this study, presumably Li2
. In the alkaline buffered media, typical Li leaching efficiency values were in the range of 1%, and the maximum value obtained after the longest leaching time of 1250 min was found to be 1.3%. Although this value was low and was determined mainly by the highly soluble Li salts (Li2
), a higher removal of Li should be avoided. Investigations in the acid solution showed that a Li removal of approximately 1.6% disintegrates the NMC secondary particles, generating significantly smaller secondary particles, which are particles consisting of only a few primary particles, and even single primary particles. Thus, the salts present on the surface contribute significantly to the cohesion of the secondary particles. This demonstrates how sensitively the chemical treatment influences the particle size, and thus, the quality of the recovered NMC. Finally, Raman spectroscopy is typically used to characterize NMC and to evaluate the quality of the recovered NMC. However, a reasonable interpretation of the Raman spectra was limited because the NMC starting material, as well as the treated NMC, were inhomogeneous. The resulting Raman spectra of the starting material and the treated NMC exhibited such a scattered range that clear statements about the quality of the treated NMC cannot be made at the present time. Further efforts—in particular, electrochemical studies on the performance of the recovered NMC—are necessary to investigate this phenomenon.