Quantifying the Impurity Distribution in Spherical Graphite: The Limitation of Flotation for Graphite Purification Explained

Abstract

Spherical graphite (SG) is a crucial raw material for the preparation of lithium-ion battery anodes. The rapid advancement of Li-ion battery materials has imposed rigorous demands on the production of ultrapure SG materials. However, SG derived from natural flake graphite (FG) via spheronization often fails to meet these quality requirements. This study investigates the physical and chemical properties of SG and the natural FG used in its production, employing techniques such as X-ray diffraction (XRD), X-ray fluorescence (XRF) analysis, and various microscopy techniques. Results reveal that FG purified via flotation retains significant impurities, and the spheronization process yields only marginal improvements in SG quality. Most impurities are distributed in the intercalation of the graphite flakes, while a smaller fraction is contributed by flotation entrainment. These distributions were visualized using FIB-SEM-EDS analysis and quantified through additional flotation tests in highly dilute solutions. This study offers a promising strategy for determining the distribution of impurities in graphite minerals and explains the limitations of flotation in upgrading graphite materials from a more microscopic perspective. Furthermore, it provides practical guidance for further SG purification using hydrometallurgical leaching techniques.

1. Introduction

Graphite, a non-metallic mineral resource, has long been utilized in various industries, including casting, refractory materials, national defense, and aerospace [1]. In recent years, its applications have gradually shifted from traditional sectors to rapidly developing strategic emerging industries, particularly the new energy battery manufacturing sector [2,3,4]. The widespread adoption and recent sustained growth of electric vehicles have increased the demand for lithium-ion batteries with high specific capacity, safety, and cost-effectiveness. As a result, graphite-based materials, including both synthetic and highly purified natural graphite, have emerged as leading candidates for producing Li-ion battery anodes due to their superior electrical conductivity and excellent charge/discharge specific capacity [5,6]. For practical applications, natural flake graphite (FG) particles undergo a mechanical spheronization process to form natural spherical graphite (SG). This transformation involves a series of sophisticated milling and classification steps, during which FG particles are bent, curled, and gradually shaped into spherical forms through constant and repetitive impacts within shaping machines [7].

Natural SG is gaining significant attention among researchers and developers today and is predicted to become the dominant state-of-the-art material for Li-ion battery anodes due to its low production cost, high capacity, and excellent cycle stability [8,9]. However, the purity of SG produced from its upstream material, natural FG, often falls short of the required purity, which typically meets or exceeds an ultra-high purity of 99.95% for practical applications. Natural FG is commonly derived from graphite ore through a froth flotation technique, a common and cost-effective method that exploits the differences in hydrophobicity between the targeted graphite mineral and unwanted gangue minerals [10].

Flotation is typically carried out in pulp solutions, where valuable mineral particles are selectively rendered hydrophobic and captured by rising air bubbles, while hydrophilic gangue minerals remain suspended and are discarded as tailings. Beyond graphite, flotation is widely used in the separation and purification of mineral resources such as pyrite [11], and chalcopyrite [12], and solid waste materials such as black mass [13].

To progressively liberate and remove gangue minerals such as mica, quartz, and pyrite from graphite while minimizing damage to the graphite flakes (since larger flake sizes typically hold higher commercial value), a multistage grinding and flotation process is commonly employed [14,15]. However, as more gangue minerals are removed, the upgrading efficiency of flotation gradually diminishes. Typically, the grade (fixed carbon content) of the final flotation concentrate of FG can be improved to about 90%–95%, but further upgrading via flotation is barely possible, as the remaining gangue minerals are interlocked within the graphite and difficult to separate through grinding and flotation [16]. Additionally, the production of SG through mechanical spheronization of FG results in minimal changes to its chemical composition. This process is further hindered by high energy consumption and low yields (only 70%–75% or less). Consequently, further purification steps are typically required after spheronization to obtain the ultra-high-grade SG necessary for Li-ion battery applications.

Currently, the purification of SG primarily relies on hydrometallurgy purification processes (e.g., the hydrofluoric acid method [17,18] and alkali-acid method [19,20,21,22]) and pyrometallurgy purification processes (e.g., the high-temperature method [23,24]). Among these, the hydrofluoric acid method is the predominant method, followed by the alkali-acid method. Although the high-temperature process is also highly efficient and eco-friendly, it is seldom used in industry due to its significantly higher operating and capital costs [8]. Despite their widespread use, hydrometallurgy methods for SG purification face several challenges, including low reaction efficiency and activation between reagents and impurities, and high dosage requirements of acids and bases due to the complex association between graphite and impurities. These problems have significantly hindered the development of the natural graphite industry in China and created obstacles to the application of graphite in high-tech industries. As the world’s largest exporter of SG and related graphite products [25,26], China holds more than 50% of its FG reserves in northeastern regions such as Jixi and Luobei. Thus, systematically studying the distribution of impurities in graphite and the mechanism of impurity transformation from FG to SG is of great practical importance.

This study aims to investigate the distribution of impurities in a natural FG concentrate from Luobei and the SG derived from it, providing theoretical insights for the subsequent chemical purification process. The chemical and physical properties of the graphite samples were analyzed using X-ray diffraction (XRD) and X-ray fluorescence (XRF) spectroscopy. The distribution of gangue minerals within the samples was examined using optical microscopy, focused ion beam scanning electronic microscopy (FIB-SEM-EDS), and dilute flotation tests. This paper represents the first investigation into the association patterns of gangue minerals in natural FG and SG from a microscopic point of view, offering an explanation of the limitations of flotation and spheronization in further improving graphite purity. It highlights the practical importance of chemical purification for graphite-based materials intended for high-end applications.

2. Materials and Methods

2.1. Materials

The FG samples used in this study were obtained from the flotation concentrate of Luobei graphite ore, and the SG samples were produced by spheronizing the FG samples. During the spheronization process, the flaky graphite particles underwent a series of milling and classification steps, gradually transforming the flaky graphite particles into spherical shapes. A scheme of graphite shape changes during the spheronization process is illustrated in Figure 1.

X-ray fluorescence (XRF) spectrum analysis (Zetium, Panalytical. B.V, Almelo, The Netherlands) and X-ray diffraction (XRD) analysis (Empyrean, Panalytical. B.V, Almelo, The Netherlands) were employed to characterize the chemical composition and mineralogical properties of both flake and spherical graphite samples. The results are presented in

Table 1. Chemical compositions of flake graphite (FG) and spherical graphite (SG) (wt%).

Composition	SiO2	Al2O3	Fe2O3	K2O	Cl	CaO	Other	LOI	FC
FG	1.78	0.54	0.83	0.21	0.011	0.014	0.015	96.60	95.13
SG	1.46	0.63	0.74	0.16	0.009	0.026	0.012	96.95	95.35

and Figure 2a, respectively. Despite some deviations, the flotation-concentrated FG and the SG produced from it exhibited highly similar chemical and mineralogical properties, with average fixed carbon (FC) contents of 95.13% and 95.35%, respectively (assayed by the National Standard GB3521–2008 [27] Graphite Chemical Analysis Method). The main impurities associated were SiO₂, Al₂O₃, Fe₂O₃, and K₂O, with other minor components accounting for less than 0.1%. The main gangue minerals associated were identified as quartz and mica. Understandably, spheronization is a physical modification process that primarily changes the shape, size, and surface morphology of the graphite samples, but not the chemical compositions. For example, a sharp particle size reduction of the SG sample was observed after the nodulizing process when compared to the particle size of the feed FG. The particle size distribution results in Figure 2b indicate that most SG particles are smaller than 25 μm, whereas most FG particles are larger than 45 μm.

More straightforwardly, Figure 3 shows the SEM images of the samples, which provide a more microscopic and intuitive comparison of the samples before (Figure 3a) and after (Figure 3b) spheronization in terms of both particle shapes and sizes. As can be seen from Figure 3a, the FG particles are stacked on top of each other, presenting irregular flakes with uneven size and thickness. The surface of the FG is smooth and flat, and a few fine granules separated from the graphite particles are clearly visible, which may be quartz, mica, or pyrite. These impurities are assumed to be mixed into the graphite samples by entrainment during flotation. Interestingly, the SG obtained after spheronization of the FG exhibits a distinct spherical or spherical-like shape (Figure 3b). Similarly, some irregular particles other than SG are observed, suggesting the presence of liberated impurities in the SG sample even after the spheronization process.

2.2. Flotation Tests

To investigate the presence of impurities in the graphite concentrate and their distribution, additional froth flotation tests were performed on both the FG and SG samples to quantify the proportion of impurities originating from the interlayers versus those from flotation entrainment. The flotation tests were conducted using an XFD-IV mechanical flotation cell (1 L) manufactured by Wuhan Exploration Machinery Co., Ltd., China (Wuhan, China). In each flotation test, a specified amount of graphite sample was first stirred in the cell at 1800 rpm with the mechanical agitation impeller. Each step of reagent adding and conditioning was maintained for 3 min, followed by aeration at a constant airflow rate of 5 L/min and bubble scraping for another 3 min. All flotation tests were carried out at a neutral pH of 7. Kerosene and terpenic oil were used as the collector and frother for flotation, respectively. Each test was triplicated, and the flotation efficiency was evaluated by the mean value of the concentrate grade (assaying the FC content) and flotation recovery.

2.3. Characterizations

A transmission/reverse-polarized optical microscope (CX40P, Sunny Instruments Co., Ltd., Ningbo, China) was used to capture microscopic images of the samples. The surface morphology of the samples was recorded using a JSM-IT300 scanning electronic microscope (SEM, JEOL, Tokyo, Japan). In addition, a focused ion beam (FIB, 30 kV, 2.5 nA, and 1 µs dwell time, ZEISS Crossbeam 540, Oberkochen, Germany) was employed at high beam currents for site-specific sputtering, milling, or etching, enabling detailed analysis of the graphite interlayers. The topography and chemical composition of these areas were further investigated using a resembled SEM and energy-dispersive X-ray spectroscope (EDS, X-Max 50).

3. Results and Discussion

3.1. Identifying the Impurity by Microscopy Techniques

The application of SG as a raw material for Li-ion battery anodes relies heavily on the proper removal of any unwanted impurities. Therefore, understanding the distribution of impurities in the products is of great interest for the development of highly efficient and economical purification techniques. For the FG collected from flotation, impurities may be present in two main forms: dissociated from graphite and interlocked with graphite. The former is easily removed via physicochemical separation methods, such as flotation under dilute solutions, while the latter is difficult to eliminate via these methods.

To investigate this issue, the distribution and content of impurities in the graphite samples were first analyzed using an optical microscope. Figure 4 presents the optical microscopy images of the samples. It can be seen that graphite is present in black under the polarizing mode (Figure 4a) and in yellow under the reflecting mode (Figure 4b). The main gangue minerals in both the FG and SG samples, identified through XRD and XRF analysis, include quartz, mica, and pyrite. Notably, while some gangue minerals can be seen freely liberated, more are sandwiched with graphite flakes, exhibiting stripe-like features. Under the microscope, the SG particles are distributed in a spherical or ellipsoidal shape, with some pyrite minerals free outside the SG being observed (Figure 4c). There may also be some gangue minerals wrapped inside the SG due to the spheronizing process, though these are not directly visible.

To gain a deeper understanding of the internal structure of the graphite samples, FIB-SEM was used to characterize the microstructure from a more internal perspective. Figure 5a shows the surface morphology of a single FG particle with a diameter of more than 100 μm. The sample stage was then tilted 52° to position the graphite flake perpendicular to the focused ion beam (Figure 5b). The ion beam was subsequently used to slice through the graphite flake, exposing its cross-section (Figure 5c). As expected, the cross-section exhibits a sandwich-like structure, with multiple graphite layers stacked on top of each other. To further determine the chemical properties of the FG, elemental mapping and analysis of selected spots were carried out on the same cross-section (Figure 5d). As shown in Figure 5e, the main constituent element is C, with only small amounts of other elements, such as Si, Fe, Al, O, and Ca. In addition, the presence of these impurities was further confirmed by the selected spot analysis (Figure 5f). Detailed elemental information is given in

Table 2. Elemental information of the selected area from the FG cross-section (wt %).

Element	C	Si	Al	Fe	Ca	O	Mg
Map sum spectrum	87.77	2.20	1.88	1.95	0.05	6.13	0.01
Spectrum 1	97.25	0.22	0.16	0.20	0.06	2.07	0.04
Spectrum 2	26.61	18.30	16.69	12.51	0.28	25.51	0.09

. Specifically, the sum spectrum shows that C is the dominant element, accounting for 87.77% of the mass fraction, while the other impurities, Si, Al, Fe, Ca, and O, show mass fractions of 2.20%, 1.88%, 1.95%, 0.05%, and 6.13%, respectively. These impurities are likely associated with quartz, pyrite, and calcite. Notably, these impurities are primarily confined within the intercalation of graphite flakes and could not be removed during flotation.

Similarly, the FIB-SEM-EDS images of SG and the elemental information are presented in Figure 6 and

Table 3. Elemental information of the selected area from the SG cross-section (wt %).

Element	C	Si	Al	Fe	Ca	O	Mg
Map sum spectrum	95.35	1.29	0.47	0.87	0.17	1.80	0.02
Spectrum 1	83.56	4.55	2.00	2.63	0.47	6.69	0.10
Spectrum 2	85.86	3.71	1.37	2.35	0.34	6.08	0.38

, respectively. A spherical SG particle with a diameter of 15 μm (Figure 6a) was first cut open by the FIB, exploring a flat cross-section (Figure 6b). Interestingly, the cross-section exhibits a core rolling feature, which can be clearly visualized in the magnified image (Figure 6c). This is thought to result from the curling and compacting of the FG flakes. No obvious granular impurities are observed inside the sphere, but some minor brighter dots are identified, which are suspected to be impurities other than graphite. Elemental mapping in Figure 6e and selected spot spectra in Figure 6f indicate that C is also the main component, while other impurities are Si, Fe, Al, O, and Ca, which are similar to those of the FG. These impurities are wrapped and interlocked inside the particles during the spheronization process. As a result, further purification by chemical processes (e.g., hydrometallurgical leaching) is needed to remove them, thereby improving the quality of the final product for high-end Li-ion battery applications.

3.2. Quantifying the Impurity Distribution by Flotation

As discussed, the FIB-SEM-EDS results have demonstrated the presence of a large proportion of impurities associated within the interlayers of the SG. Moreover, there are still some observable liberated impurities, probably resulting from entrainment during FG flotation and remaining unaffected during the spheronization process. Entrainment leads to the collection of a portion of impurity minerals that have been completely dissociated from the target minerals, ultimately leading to the grade deterioration of the graphite concentrate. It is widely acknowledged that the entrainment of gangue minerals is strongly influenced by the solid concentration of the flotation pulp, with severity typically increasing as the solid concentration rises [28,29]. Some of the literature suggests that when the flotation concentration is low enough, the chances of entrainment become minimal or even eliminated [28,30,31]. In light of this, further flotation experiments under relatively low solid concentrations were designed and performed to quantitatively assess the percentages of impurities contributed by entrainment versus those associated with graphite interlayers.

Figure 7 shows the flotation of FG and SG in relation to varying solid concentrations in the pulp, with a comparison of three different collector dosages. As expected, the FC content of the FG concentrate gradually increased with the decrease of solid concentration, regardless of the collector dosage (Figure 7a). In contrast, the recovery exhibited a constant decrease (Figure 7b). A similar trend was observed in the flotation of SG (Figure 7c,d). Specifically, when the solid concentration dropped to 1%, the FC content of both FG and SG flotation concentrates reached their maximum values. This suggests that the effect of entrainment on flotation is minimal, and most of the impurities present in the flotation concentrate are assumed to be those not dissociated from graphite particles. These impurities are carried into the flotation concentrate alongside the collection of graphite particles. Accordingly, it can be inferred that the ratio of dissociated impurities to interlayer impurities in both SG and FG is roughly similar.

For example,

Table 4. Flotation behaviors and impurity distribution of FG and SG (%).

Products	Before Flotation		After Flotation		Impurity Distribution
	FC, %	Ash Content, %	FC, %	Ash Content, %	In Interlayer, %	By Entrainment, %
FG	95.13	4.02	95.82	3.38	84.08	15.92
SG	95.35	3.85	96.41	2.84	73.77	26.23

presents the FC content and ash content of the samples before and after flotation, under the condition of 1% solid concentration and a low collector dosage of 73 g/t. Under such a dilute pulp and low collector dosage, it is anticipated that most of the graphite will be recovered by true flotation, while the liberated gangue minerals will cease being collected. Specifically, the FC content of the FG increased from 95.13% to 95.82% and the ash content decreased from 4.02% to 3.38% after flotation. The increase in FC content and the decrease in ash content are attributed to the minimized collection of fully liberated gangue minerals that were previously collected by entrainment. In other words, the remaining ash contents originate from impurities associated with graphite and contained within the interlayers. Similarly, the flotation results for the SG exhibit a similar pattern, with the FC content increasing from 95.35% to 96.41% and the ash content decreasing from 3.85% to 2.84%. Moreover, due to the high hydrophobicity of both FG and SG, it is reasonable to assume that they can be effectively collected through dilute flotation. Roughly, an estimation of the proportion of interlayer impurities and entrained impurities can be made by analyzing the ash content in the FG and SG before and after the dilute flotation. As such, it can be calculated that approximately 84% of interlayer impurities are present in the FG, while 74% are found in the SG. This observation of reduced interlayer impurities in SG probably resulted from the impact of spheronization on breaking down the FG particles, leading to further liberation of gangue minerals from the interlayers.

Furthermore, the above findings highlight the limitation of flotation in purifying graphite minerals, particularly when the FC content exceeds 95%. While there may be some minor entrainment of liberated gangue minerals, most of the impurities are present within the graphite interlayers. As a result, in order to avoid excessive destruction of graphite flakes, it is generally not feasible to achieve a further improvement in the FC content of graphite concentrate through regrinding–flotation. Rather, chemical purification is necessary to further extract these impurities.

In summary, the foregoing analysis indicates that most of the impurities in the graphite concentrate are situated within the graphite interlayers, making them difficult to dissociate and separate from the graphite via physical milling and flotation processes. This was corroborated by our flotation tests in dilute solutions of 1% solid concentration, which only resulted in a slight increase in the grade of flotation concentrate. Theoretically, under such dilute pulp conditions, only highly hydrophobic graphite particles could be effectively floated, and the ash content assayed therein could be attributed solely to these associated and interlocked impurities.

4. Conclusions

This paper investigated the presence and distribution of impurities in SG particles produced from natural FG via spheronization. Characterization techniques, including XRD, XRF, optical microscopy, SEM, and FIB-SEM-EDS, revealed that despite achieving high grades up to 96%, significant amounts of impurities persist in both FG and SG samples, comprising liberated and associated gangue minerals. Their presence strictly hinders the potential of SG materials for use in high-end applications, such as Li-ion battery anodes. Quantitatively, the flotation studies conducted under extremely dilute conditions indicated that more than 84% of the impurities in FG are associated with the graphite particles, while only a small proportion originates from flotation entrainment. Similarly, for SG, 74% of the impurities remain interlocked after spheronization. These sandwich-like impurities are unlikely to be liberated from the graphite particles via conventional flotation, even under very dilute pulp conditions. Consequently, chemical purification procedures are essential for the removal of these impurities. This groundbreaking study represents the first to visually demonstrate the presence and accurately quantify the proportion of impurities within the interlayers of graphite particles. These findings establish a theoretical basis for elucidating the limitations of traditional flotation in achieving ultrapure graphite and highlight the necessity of employing chemical processes to further purify graphite materials for Li-ion battery applications.