Influence of Flow-Gas Composition on Reaction Products of Thermally Treated NMC Battery Black Mass

Abstract

The recycling of lithium-ion batteries (LIBs) is becoming increasingly important regarding the expansion of electromobility and aspects of raw material supply. Pre-treatment and liberation are crucial for a sufficient recovery of all relevant materials from LIBs. Organic removal and phase transformations by thermal pre-treatment are beneficial in many respects. This study deals with the influence of flow-gas composition on reaction products and water-based lithium recovery after thermal treatment. Therefore, a spent NMC black mass was thermally treated at 610 °C in a moved bed batch reactor under an N₂ atmosphere and mixtures of N₂ with 2.5% and 5% O₂. Since the phase transformation of the lithium content to Li₂CO₃ is targeted for water leaching, a treatment under a CO₂ atmosphere was studied as well. The resulting off-gas was analyzed by FTIR, and the black mass was characterized by XRD. Afterward, water washing of the black mass was carried out for selective lithium recovery. The gained lithium product was analyzed for the purity and phases present. The addition of O₂ resulted in reduced reduction reactions of lithium metal oxides and lower Li-yields in the water leaching compared to the other two atmospheres. In the case of CO₂, the formation of Li₂CO₃ is favored compared to LiF, but the Li-yield of 56% is comparable to N₂ treatment.

1. Introduction

The ongoing electrification of our society leads to rising demands for energy storage systems. Therefore, LIB production is rapidly increasing. This leads to a high demand for battery raw materials such as cobalt, nickel, manganese, graphite, copper, and in particular lithium [1,2]. To ensure the availability of raw materials and, at the same time, meet the requirements of a circular economy, the recycling of used batteries and production scrap is essential [3]. Thermal pre-treatment of battery scrap before or after entering the first mechanical recycling steps offers several advantages, as described in the literature [4,5,6,7,8,9,10]. The main objective is the separation and evaporation of organics, originating from electrolytes, binders and separators [7,11]. In combination with the targeted phase transformation of the contained metal oxides [12], the thermal treatment process leads to the better leaching behavior of the black mass in the following hydrometallurgical steps [8,11] and allows for early-stage lithium recovery (ESLR) by a combination of thermal treatment and water leaching [5,9,13,14,15,16,17,18,19,20]. The thermal-induced phase transformation of the NMC cathode material enables the formation of water-soluble Li₂CO₃ as a reaction product. Afterward, the Li₂CO₃ can be recovered by the selective water leaching of the black mass fraction. This concept is promising since recycling lithium is challenging [5,9]. In common hydrometallurgical process routes, precipitation of lithium salts is the final step. Parts of the lithium are already lost in previous cementation and precipitation steps of the other battery metals [5,9]. In case of pyrometallurgical battery recycling, lithium is transferred to the slag phase [21,22,23] or flue dust [22,24,25,26,27,28,29]. This requires subsequent mechanical and hydrometallurgical processing of the slag for lithium recovery, which is energy and chemical intensive. With further improvement, the reported concept of ESLR could overcome those drawbacks. Previous studies on spent NMC LIB shredders or black mass achieved lithium recovery yields of 60–64% after thermal treatment at around 600–700 °C [5,14,15,30]. Therefore, a deeper understanding of the process and the influence of its parameters on lithium recovery is necessary for process improvement.

The previous studies described below investigated the behavior of battery materials during pyrolysis under an inert atmosphere and incineration with air. Under an inert atmosphere, an increasing reduction in the contained NMC material with rising temperature was observed [12,16,17,30]. This leads to water-soluble Li₂CO₃ formation as a reaction product [14,15,30]. In contrast, Lombardo et al. found that after the incineration of NMC battery material with air, no reduction in the metal oxides could be observed in XRD analysis in the investigated temperature range from 450 to 650 °C [6]. Consequently, the ESLR process would not be promising in this case. Regarding organic removal and off-gas generation during incineration and pyrolysis, few studies are reported in the literature for different input materials. In the case of incineration of cathode materials, major products are reported to be CO₂, volatile hydrocarbons and H₂O [6,31]. In the case of incomplete combustion, as for the pyrolysis of whole battery cells, CO is detected as well [32]. Additionally, some studies deal with fluorine emissions [6,31,32], but no research was found regarding the influence of different oxygen concentrations in the flow-gas on off-gas and phase composition in the black mass.

In the case of CO₂ flow gas, to the best of the author’s knowledge, not much has been reported in the literature to date. Park et al. [33] investigated the thermal treatment of pure NMC 811 and graphite mixtures under CO₂ and gained higher Li₂CO₃ recovery rates by water leaching compared to inert N₂ treatment. Schwich et al. [34] reported a leaching efficiency of 60% for thermally treated NMC black mass under a CO₂ atmosphere, but without focus on the behavior or influence of CO₂ during the thermal treatment compared to other flow-gas compositions. For a better understanding of the process and in order to develop further optimization options for thermal treatment, the present study investigates the influence of different flow-gas compositions on both the off-gas and the black mass achieved, as well as the effects on the ESLR process. An inert N₂ atmosphere, with mixtures of N₂ with up to 5% O₂, and a pure CO₂ atmosphere are investigated comparatively at 610 °C. The trials were carried out in an oscillating tube furnace to ensure maximized solid–gas contact between black mass and flow-gas. The maximum tolerable O₂ amount in the process atmosphere to ensure the conversion of Li to Li₂CO₃ as completely as possible is to be identified within this work for the first time. Information in this regard is also of interest for plant engineering and construction. Additionally, the promising results of Park et al. [33], who gained higher Li recovery yields by thermal treatment under a CO₂ atmosphere compared to N₂ from a mixture of pure NMC 811 and graphite, are transferred and investigated for an end-of-life battery black mass containing further potential interfering elements, such as fluorine. Since the process is carried out in a moved bed reactor with enhanced solid–gas interaction, higher Li₂CO₃ yields are expected compared to previous studies with CO₂ addition. The addition of O₂ to the process suggests increased CO₂ formation in the exhaust gas due to combustion reactions as well. Accordingly, in this study it is to be evaluated whether the oxidation behavior of the NMC oxides, or the favored formation of Li₂CO₃ under CO₂ atmospheres expected from the literature [33,34], prevails as a function of the O₂ addition. To assess carbonate formation in the context of the thermal pre-treatment of battery materials, this study is the first, to the authors’ knowledge, to determine the inorganic carbon content after water leaching.

2. Materials and Methods

The input feed for the thermal treatment trials was an NMC 622 LIB black mass generated from an industrially inert shredder process under an N₂ atmosphere, followed by drying below 80 °C. Separation of the black mass from the shredded material was carried out by sieving at 0.5 mm at TU Freiberg.

Table 1. Chemical composition of the black mass.

Element	Al	Cu	Co	Li	Mn	Ni	P	C	F
wt.%	2.02	1.83	5.30	3.21	5.09	14.6	0.55	39.90	2.50

provides the chemical composition of the black mass fraction, analyzed by the ICP-OES (Spectro CIROS Vision, Spectro Analytical Instruments GmbH, Kleve, Germany) combustion method for carbon analysis (ELTRA CS 2000, ELTRA GmbH, Haan, Germany) and ion chromatography (811 Compact IC pro, Deutsche Metrohm GmbH & Co. KG, Filderstadt, Germany).

The thermal treatment trials were conducted in an oscillating quartz-tube furnace (TSO 11/400, Carbolite Gero GmbH & Co. KG, Neuhausen, Germany). Therefore, 80 g active mass per trial, and 100 g in case of CO₂ treatment, was fed into the quartz tube. The material was heated up simultaneously with the furnace with a heating rate of 300 °C/h to 610 °C, followed by a 1 h holding time. The temperature was measured in the center of the tube with a type-N thermocouple. The tube was oscillating during the trial in a range of 315 ° and a frequency of 7 full oscillations per minute. During the heating and holding time, the different investigated flow-gas compositions were inserted with a flow rate of 4 L/min:

• N₂ (trials in triplicate);
• N₂ + 2.5% O₂ (trials in duplicate);
• N₂ + 5% O₂ (trials in duplicate);
• CO₂ (trials in duplicate).

After the holding time, the heating power was turned off, and the flow-gas was switched to N₂. Oscillation and N₂ gas flow were turned off when the temperature dropped below approx. 200 °C. The resulting off-gas, except for the trials under a CO₂ atmosphere, was continuously analyzed by an FTIR (CX4000 FTIR gas analyzer, Gasmet Technologies Oy, Vantaa, Finland) and additional O₂ (Oxygen Analyzer PMA 10, M&C TechGroup, Ratingen, Germany) and H₂ (CONTHOS 3—TCD, LFE GmbH & Co. KG, Bruchköbel, Germany) measuring units for its composition. For exhaust cleaning, the gas flowed through a two-stage scrubbing unit first with NaOH solution and second with water, followed by a catalytic post-combustion. The <90 µm fraction of the thermally treated black mass was analyzed by XRD (MiniFlex, 40 kC, Cu-tube, Rigaku) and SEM EDX (S3700N Hitachi, Quantax 200 detector, 123 eV MnKalpha).

After thermal treatment, water leaching of 20 g black mass per trial was carried out with 500 mL deionized water in a glass backer. The solution was stirred for 90 min at room temperature at 200 rpm. During subsequent filtration, the filter cake was washed with 200 mL of deionized water and afterward dried at 80 °C for 24 h. The filtrate was analyzed for Li, Al and P by ICP-OES and by TOC analyzer for organic and inorganic carbon in solution (multi N/C 2100/2100S, Analytik Jena GmbH, Jena, Germany). Fluorine determination was carried out by ion-selective electrodes (Methrom Titrando 888, Deutsche Methrom GmbH & Co. KG, Filderstadt, Germany). Analysis of the filter cake was carried out similarly to the input material for the elements Li, Al, F, P and C.

The filtrate was vaporized by boiling the solution to recover a lithium salt product. The purity and phase composition of residual salt were analyzed by ICP-OES, combustion method, IC and XRD, similar to the previous samples. The whole process is summarized in Figure 1.

3. Results and Discussion

The following sections provide the results of the thermal treatment trials, including off-gas and black mass characterization, as well as the outcomes of water-leaching trials with the different thermally treated black masses.

3.1. Off-Gas Analysis

The analysis of resulting off-gas from the pyrolysis trials was carried out for a deeper understanding of the influence of different oxygen partial pressures in the flow-gas on the ongoing reactions during the process. In general, several different gases were detected, such as CO, CO₂, H₂O; hydrocarbons; other organics such as alcohols and aromatic compounds; electrolyte compounds EC, DMC, DEC, EMC and HF; and POF₃ emissions, as shown in Figure 2.

Because of this complex off-gas composition, exact analysis is challenging and limited. During some trial periods, the gas is highly concentrated, which results, together with interferences/overlaps of different compounds, in high intensities of the measured spectrums. This, in turn, leads to inaccuracies in the measured component concentrations. Moreover, not all evaporating gases can be identified due to a leak of reference data. This could be the case, e.g., for some fluorinated hydrocarbons, reported to originate from PVDF decomposition [35]. Nevertheless, the presence of the addressed gases can be determined, but with inaccuracies of the measured concentrations. However, the qualitative curve progression and their comparison within the different trials is possible and provides information about different ongoing reactions as regards their dependence on the oxygen partial pressure.

The comparison of CO₂ and CO generation during the trials between the different oxygen partial pressures, presented in Figure 3, shows similar minima and maxima and temperature timings but different amounts of gas release. At around 150 °C, the first peak of CO₂ generation is detectable with comparable concentration between all trials. It is attributed to the decomposition of the conducting salt and the solid electrolyte interface, according to Equation (1) [32,36].

$$(\mathrm{C}{\mathrm{H}}_2\mathrm{O}\mathrm{C}{\mathrm{O}}_2\mathrm{L}\mathrm{i}{)}_2\to {\mathrm{L}\mathrm{i}}_2{\mathrm{C}\mathrm{O}}_3+{\mathrm{C}}_2{\mathrm{H}}_4+{\mathrm{C}{\mathrm{O}}_2+\frac{1}{2}\mathrm{O}}_2$$

(1)

The second peak of CO₂ release occurs at approx. 300 °C but differs in the concentration of CO₂ by the factor of 1.9 (N₂ vs. 2.5% O₂) and 2.8 (N₂ vs. 5% O₂). Further characteristic peaks and plateaus are detectable at around 400 °C and 500 °C and within the start of the holding time at 610 °C. The difference in CO₂ concentration is also rising with temperature; e.g., for the third CO₂ peak at 400 °C, the factors are 3.4 (N₂ vs. 2.5% O₂) and 8.1 (N₂ vs. 5% O₂). The peaks are attributed to the organic decomposition from the battery material. The main electrolyte compound of the investigated material is EC, with a boiling temperature of 248 °C [37]. Therefore, its release and also decomposition is detected from this temperature. Between 400 and 500 °C, the decomposition of separators and binders takes place [31]. Because the detectable peaks during the gas release occur at similar temperatures for all considered atmospheres, it can be concluded that the decomposition temperatures of the organic compounds are not shifted. Therefore, to reach complete organic removal, the same maximum temperature is necessary. In the tests with 0% and 2.5% O₂, an anomaly occurs in the gas measurement at the hold time onset. There is a short but strong drop in concentration followed by a rapid increase. This drop is due to a brief interruption of the gas supply by the furnace control unit and not due to chemical reactions.

In the case of CO, the qualitative curve progression is similar between the trials but differs in quantitative results as well. In general, higher O₂ partial pressure results in higher amounts of released combustion products: CO₂ and CO. As a result, lower hydrocarbon concentrations are detected, as presented in Figure 3d. The shape of the curves differs between the trials, so the maxima of hydrocarbons occur at different temperatures for different O₂ concentrations. In particular, the components CH₄ and C₂H₄, which make up the largest share, occur in lower concentrations due to the partial combustion resulting in CO₂, CO and H₂O, according to Equations (2) and (3) [38].

$${\mathrm{C}}_x{\mathrm{H}}_y+\left(\mathrm{x}+\frac{\mathrm{y}}{4}\right){\mathrm{O}}_2\to \mathrm{x}{\mathrm{C}\mathrm{O}}_2+\frac{\mathrm{y}}{2}{\mathrm{H}}_2\mathrm{O}$$

(2)

$${\mathrm{C}}_x{\mathrm{H}}_y+\left(\frac{\mathrm{x}}{2}\right){\mathrm{O}}_2\to \mathrm{x}\mathrm{C}\mathrm{O}+\frac{\mathrm{y}}{2}{\mathrm{H}}_2\mathrm{O}$$

(3)

The three maxima of hydrocarbon release under an inert atmosphere, attributed to the different phases of pyrolysis processes, as described and discussed in another study by the authors [30], are not present in the case of O₂ addition. With rising temperature, the combustion reactions are favored. The reactions of O₂ with solid carbon take place during the holding time since the CO₂ concentration remains at a high level during the holding time, even though no further hydrocarbons are released in this temperature range, also in an inert atmosphere. This is confirmed by the XRD characterization of the black mass in Section 3.2, but it cannot be identified if the CO₂ is a reaction product from the combustion of formed pyrolysis coke or graphite. The decreasing CO₂ concentration over the holding time under an inert atmosphere indicates the termination of both organic decomposition and reduction reactions of metal oxides.

Regarding HF release, a general correlation between HF concentration and O₂ partial pressure can be derived from the FTIR analysis: With increasing O₂ addition, the HF concentration in the off-gas decreases (see Figure 2). Diaz et al. [32] found a similar correlation by comparing treatment under air and N₂. However, since the total off-gas volume could not be measured in the present study, it cannot be determined whether total HF emission is lower or if its concentration is diluted.

3.2. Black Mass Characterization

Characterization of the produced black mass by SEM-EDX and XRD measurement allows for further conclusions regarding chemical reactions during thermal treatment. The comparison of SEM images between untreated and black mass treated with different oxygen concentrations, presented in Figure 4, indicates structural changes regarding the NMC particles.

NMC particles (bright, light grey) from thermally treated samples look more spending than those untreated, especially with increasing O₂ content during thermal treatment. The shadows around the individual particles caused by the binder [39] are also reduced by thermal treatment, particularly by the addition of O₂. However, binder residuals around the NMC particles can be detected in all samples. Whether the approx. 10 µm large NMC particles have decomposed to smaller particles due to the thermal treatment cannot be accurately estimated from the SEM images.

In all samples, Al can be detected by EDX-spot analysis on the NMC particles in shares of 1.5–6 wt.%. Since the black mass originates from NMC 622 cells, this indicates an alumina coating of the particles, as reported in the literature [40,41].

In the case of thermal treatment under a CO₂ atmosphere, black mass particles, presented in Figure 5, look similar to inert treated material, with binder residues detectable around NMC particles. In contrast to the other SEM images, aluminum content, measured by EDX-spot analysis, is higher. Elemental analysis of the large particle in the lower right corner of the image provides an Al share of 94 wt.%. Its origin cannot be identified, but the partial smelting and oxidation of aluminum foil or the agglomeration and sintering of single particles could be possible explanations.

The XRD analysis with HighScore [42] of the different treated black masses, given in Figure 6, shows different phases depending on the flow-gas composition during thermal treatment. In the case of inert treatment under N₂, the lithium-nickel-manganese-cobalt-oxides (LNMCO) from the untreated black mass are no longer detectable. They were reduced to NiO, metallic Ni, according to the exemplary reactions (4)–(7) derived from [12,43], and, in contrast to a previous study with a whole battery shredder in a fixed-bed reactor [30], to LCO. Since reflexes of different metal oxides (e.g., MnO, NiO, CoO) overlap, they cannot be clearly distinguished in the XRD pattern, but their existence is assumed. Another product of the reduction is Li₂CO₃ (see Equation (8), the reflexes of which are identified. In contrast, with the addition of O₂ during thermal treatment, Li₂CO₃ is not detectable by XRD. The lithium metal oxides are less reduced with rising O₂ concentration so that LMNO, LMO and LCO are still present in the black mass, and with 5% O_2, no metallic Ni is detectable. Investigations of Lombardo et al. [6,44] concluded that incineration under air results in no change in the phase composition of the black mass compared to the untreated sample with respect to the cathode material. The present study shows additionally that 2.5% O₂ already inhibits the reduction reactions but not to the same extent as incineration under air. Potential reaction partners such as hydrocarbons are no longer available to the same extent due to combustion reactions. Moreover, a comparison of the intensity of graphite reflexes between the different O₂ concentrations indicates the partial combustion of it by O₂ addition. This is confirmed by CO₂ formation during the holding time of the experiments, described in Section 3.1.

$$8\mathrm{C}{\mathrm{H}}_4+\mathrm{L}\mathrm{i}\mathrm{N}\mathrm{i}{\mathrm{O}}_2\to {\mathrm{C}\mathrm{O}}_2+4\mathrm{L}{\mathrm{i}}_2\mathrm{O}+8\mathrm{N}\mathrm{i}\mathrm{O}+2{\mathrm{H}}_2\mathrm{O}$$

(4)

$$2\mathrm{L}\mathrm{i}\mathrm{N}\mathrm{i}{\mathrm{O}}_2+\mathrm{C}\mathrm{O}\to {\mathrm{C}\mathrm{O}}_2+\mathrm{L}{\mathrm{i}}_2\mathrm{O}+2\mathrm{N}\mathrm{i}\mathrm{O}$$

(5)

$$4\mathrm{L}\mathrm{i}\mathrm{N}\mathrm{i}{\mathrm{O}}_2+\mathrm{C}\to {\mathrm{C}\mathrm{O}}_2+2\mathrm{L}{\mathrm{i}}_2\mathrm{O}+4\mathrm{N}\mathrm{i}\mathrm{O}$$

(6)

$$\mathrm{N}\mathrm{i}\mathrm{O}+\mathrm{C}\mathrm{O}\to {\mathrm{C}\mathrm{O}}_2+\mathrm{N}\mathrm{i}$$

(7)

$$3\mathrm{L}{\mathrm{i}}_2\mathrm{O}+3\mathrm{C}{\mathrm{O}}_2\to 3\mathrm{L}{\mathrm{i}}_2\mathrm{C}{\mathrm{O}}_3$$

(8)

In the case of thermal treatment under a CO₂ atmosphere, LNMCO is decomposed to LCO, LMCO and metallic Ni. Similar to inert N₂ treatment, targeted Li₂CO₃ is identified in the XRD pattern, but since LMCO is detectable in contrast to the inert-treated sample, a weaker reduction is assumed. In the literature, favored CO production due to the Boudouard reaction under a CO₂ atmosphere at 600 °C in the presence of graphite was reported [33]. Due to verifiable LMCO and LCO content in the sample, it is more likely that in the case of organics containing EoL black mass, the thermochemical equilibrium of organic cracking with a CO₂ product (see Section 3.1) is shifted and results in a more endothermic process than under an inert atmosphere. Therefore, decomposition reactions could be hindered. This explanation will be further investigated in following studies.

3.3. Water Leaching

Water-leaching trials were carried out to recover lithium from the black mass and obtain an enhanced understanding of the reactions taking place during the thermal treatment. Therefore, leaching efficiencies of Li, Al and F were calculated based on Equation (9).

$${\mathrm{L}\mathrm{E}}_{{\mathrm{x}}_{\mathrm{i}}}=\frac{{\mathrm{x}}_{\mathrm{i}} \mathrm{i}\mathrm{n} \mathrm{s}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \left[\mathrm{g}\right]}{{\mathrm{x}}_{\mathrm{i}} \mathrm{i}\mathrm{n} \mathrm{s}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n} \left[\mathrm{g}\right] +{\mathrm{x}}_{\mathrm{i}} \mathrm{i}\mathrm{n} \mathrm{f}\mathrm{i}\mathrm{l}\mathrm{t}\mathrm{e}\mathrm{r}\mathrm{c}\mathrm{a}\mathrm{k}\mathrm{e} \left[\mathrm{g}\right]}·100\mathrm{\%}$$

(9)

$${\mathrm{L}\mathrm{E}: \mathrm{l}\mathrm{e}\mathrm{a}\mathrm{c}\mathrm{h}\mathrm{i}\mathrm{n}\mathrm{g} \mathrm{e}\mathrm{f}\mathrm{f}\mathrm{i}\mathrm{c}\mathrm{i}\mathrm{e}\mathrm{n}\mathrm{c}\mathrm{y}; \mathrm{x}}_{\mathrm{i}}: \mathrm{C}\mathrm{o}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{d}\mathrm{e}\mathrm{r}\mathrm{e}\mathrm{d} \mathrm{E}\mathrm{l}\mathrm{e}\mathrm{m}\mathrm{e}\mathrm{n}\mathrm{t}(\mathrm{L}\mathrm{i}; \mathrm{F}; \mathrm{A}\mathrm{l})$$

The resulting leaching efficiencies for Li are presented in Figure 6. As expected from previous research [30] and the literature [14,15], inert thermal treatment results in better Li leaching efficiency than from untreated black mass due to the described reductive reactions of lithium metal oxides in Section 3.2. From untreated material, 13.5% of Li is recovered by water leaching. The leaching efficiency is increased to 55.7% by inert thermal treatment, which is lower than the 62% obtained in previous research with a complete shredder fraction [30] or other studies from the literature with coated foil fractions and separators [14]. This could be explained by the lower organic content of black mass compared to the whole shredder fraction; e.g., Rouquette et al. [14] found a positive impact of the presence of a separator material during thermal treatment on Li recovery by water leaching. When treating a black mass fraction without a separator under an inert atmosphere, they reached an Li recovery of 35% after water leaching, compared to 61% in the case of separator addition [14]. The Li recovery rate in this study is located between those two values, which is probably due to the different organic contents of the input materials. This will be further investigated in following studies.

The comparison of Li leaching efficiency between samples from N₂ and N₂ + O₂ atmosphere treatments shows a negative impact of O₂ addition during thermal treatment on Li recovery by water leaching (see Figure 7). The described inhibited reduction reactions of lithium metal oxides described in the previous chapters result in less formation of water-soluble Li₂CO₃. For this reason, the leaching efficiency decreased from 55.7% to 36.6% and 27.4%. The decrease in inorganic carbon concentration in the solution with rising O₂ addition during thermal treatment confirms this explanation.

In the case of thermal treatment under a CO₂ atmosphere, the Li leaching efficiency of 56.2% is similar to the results of N₂ treatment. However, the amount of dissolved inorganic carbon in the solution is 614.5 mg/L and higher than for N₂ treatment with 497.3 mg/L. This allows for the conclusion of higher dissolved carbonate salt amounts. The comparison with the behavior of fluorine leachability, shown in Figure 8, supports this explanation: The leaching efficiency of F, as well as the F concentration in solution, is lower for black mass from CO₂ thermal treatment than from pyrolysis under N₂. This could be due to a more preferential formation of Li₂CO₃ instead of LiF, because of the shift of the equilibrium of reaction (10) [45,46] during thermal treatment under a CO₂ atmosphere. Fluorine leaching efficiency is also lower for black mass from thermal treatment with O₂ addition compared to N₂ treatment. A decreased reduction in lithium metal oxides results in less Li₂CO₃ formation and therefore the reduced possibility of a reaction of Li₂CO₃ with released HF, according to Equation (10).

$$\mathrm{L}{\mathrm{i}}_2\mathrm{C}{\mathrm{O}}_3+2\mathrm{H}\mathrm{F}\to 2\mathrm{L}\mathrm{i}\mathrm{F}+\mathrm{C}{\mathrm{O}}_2+{\mathrm{H}}_2\mathrm{O}$$

(10)

Moreover, as described in Section 3.1, HF concentration in the off-gas lowers with rising O₂ addition, resulting in the decreased reaction activity of HF.

Besides F, Al is the main impurity in the water-leaching process. The co-leaching behavior is presented in Figure 8. In the case of the untreated material, the leaching efficiency of Al is 0.3%. It rises to 4.5% and up to 6.3% for N₂ and N₂ + O₂ treatment. Between the black masses from thermal treatments under different N2 + O₂ mixtures, no coherence of Al leaching efficiency can be derived. In contrast, thermal treatment under CO₂ leads to higher leaching efficiencies of Al. This can be attributed to the higher pH value of the solution due to the larger amount of Li₂CO₃ that causes Al corrosion as derived from the Pourbaix diagram of Al [47]. Although the formation of AlF₃ is possible during pyrolysis [15,46], the increased solubility of this salt is not assumed due to the lower F content in the solution compared to N₂ trials.

After leaching, the precipitation of the lithium salt product was carried out by the complete evaporation of water by boiling. An additional XRD analysis of the generated salts, presented in Figure 9, confirms the described relationships between Li, F and IC analysis. All samples mainly consist of Li₂CO₃ and LiF, with dominating carbonate reflexes. As expected from the leaching results and black mass analysis, the semi-quantitative carbonate shares of the samples originating from treatment with O₂ addition are lower than for N₂ and CO₂ treatment. In addition, unidentifiable phases occur in the two samples with O₂ addition. Therefore, the given LiF and Li₂CO₃ shares are semi-quantitative and normalized values. In the case of CO₂ thermal treatment, the largest share of Li₂CO₃ was achieved, as expected from the IC and F analysis in solution. The Al impurities are not detectable in XRD analysis due to low concentrations.

The elemental impurity concentrations of the salts were determined for the two process parameters with the best lithium leaching efficiencies (N₂ and CO₂ thermal treatment). The averaged values of the retries are given in

Table 2. Elemental concentration of impurities in Li salts.

Atmosphere	Al	F	P	Mn	Ni	Co	Cu
	wt.%			ppm
N2	1.00 ± 0.26	9.99 ± 0.21	0.20 ± 0.02	59	110 ± 29	89 ± 34	<50
CO2	1.47 ± 0.03	7.22 ± 0.28	0.09 ± 0.003	<50	71 ± 5	54 ± 4	<50

.

The main impurities in the Li salt are Al and F. The described different leaching behaviors of Al and F for different pre-treatment atmospheres is confirmed. The Al content in the salt generated from the thermally treated black mass under an N₂ atmosphere is lower, but therefore F content is higher. Cu and Mn content were, except for one sample, below the verification limit of ICP-OES. Ni and Co impurities of 54–110 ppm were detected but may originate from small particles that have not been filtered when separating the black mass fraction from the leach solution, since the metal oxides are not soluble in water [48]. Although Co can be dissolved in the basic environment, Ni cannot [46], and their co-leaching has not yet been reported in the literature. Nevertheless, no pure Li₂CO₃ product is generated, so the further processing of the product or adjustment of the leaching parameters is necessary in future studies, as described in the literature as well [14].

The presented results show a significant influence of the selected gas atmosphere on the products of the thermal treatment and, in particular, on lithium recovery. Based on the results, CO₂ is considered as the favored atmosphere regarding targeted Li₂CO₃ formation. The suppressed formation of LiF leads to reduced F impurities in the Li salt product. Although the Al impurity increases, this is also to be expected with further Li yield increases due to the basic pH value of the solution. Since even low O₂ concentrations of 2.5% lead to a decreased reduction in the NMC oxides, an O₂-free atmosphere should be preferred in case of water-based Li recovery. Future studies will focus on a more complete reduction in the NMC oxides and the prevention of side reactions of Li with other battery elements.

4. Conclusions

The influence of different flow-gas compositions in thermal treatment processes of spent NMC black mass on the gaseous and solid products and the following water-based lithium recovery step were investigated in this study. The addition of O₂ to the flow-gas leads to the partial combustion of organics contained in the black mass by generating higher CO₂ and CO concentrations in the off-gas. Their concentrations increase with increasing O₂ addition. In this context, the concentration of hydrocarbons decreases.

Regarding the black mass products, rising O₂ addition and the changed off-gas composition lead to less reduction reactions of lithium metal oxides in the black mass, compared to inert N₂ treatment. In contrast to the combustion under air described in the literature [46], reduction reactions still take place in the investigated concentration range of O₂ addition but are incomplete even for the lowest investigated O₂ addition of 2.5%. This has a significant influence on the water-leachability of contained lithium since the transformation to Li₂CO₃ is dependent on the reduction reactions. Therefore, the leaching efficiency of lithium is decreased from 55.7% to 36.6% (2.5%O₂) and 27.4% (5%O₂).

Thermal treatment under a CO₂ atmosphere resulted, according to XRD analysis, in less complete lithium metal oxide decomposition and a more endothermic process compared to N₂ treatment, but similar Li leaching efficiency of 56.2% in the water-leaching step. A difference is the higher and beneficial IC content in solution as well as the lower leaching efficiency of F compared to N₂-treated material. The assumed suppression of LiF formation needs to be investigated in follow-up studies. Since Li₂CO₃ formation is favored for Li recovery, a CO₂ atmosphere offers a high potential to condition the LIB material for subsequent metal recovery. To further enhance the Li recovery, a follow up study is planned with whole LIBs shredder fractions and parameter adjustment to create the most reducing conditions as possible.