Conductive Additives Effects on NCA–LFMP Composite Cathode in Water-Based Binder for High-Safety Lithium-Ion Batteries

Abstract

Lithium nickel–cobalt–aluminum oxide (NCA) is a promising cathode material for lithium-ion batteries due to its high energy density of more than 274 mAh/g. However, thermal runaway inhibits its practical applications. Lithium ferromanganese phosphate (LFMP), due to its olivine structure, can effectively stabilize the surface stability of NCA and reduce the exothermic reactions that occur during thermal runaway. LFMP can also inhibit cathode expansion and contraction during charging and discharging. To improve the conductivity of an NCM–LFMP composite electrode, three different conductive additives, namely carbon black, carbon nanotubes (CNTs), and graphene, were introduced into the electrode. Finally, battery safety tests were conducted on 1.1 Ah pouch cells fabricated in the present study. The energy density of the NCA–LFMP 1.1 Ah lithium-ion pouch cells with only 0.16% CNT content reached 224.8 Wh/kg. The CNT–NCA–LFMP pouch cell was also the safest among the cells tested. These results provide a strategy for designing high-energy-density and safe pouch cells for energy storage device applications.

1. Introduction

Lithium-ion batteries (LIBs) have become widely used as power sources in consumer and medical devices, energy storage systems, remotely controlled devices, and electric vehicles. This is primarily due to their high energy density [1,2,3]. The positive electrode, or cathode, of an LIB primarily consists of lithium metal oxides, such as lithium cobalt oxide (LCO), lithium nickel manganese cobalt oxide (NMC), and lithium nickel cobalt aluminum oxide (NCA) [4]. Due to cost and energy density tradeoffs, LCO is preferred for use in mobile phones and laptops, whereas NMC and NCA are used in electric vehicles. Increasing the energy density of electric vehicle batteries is essential for producing cars that can travel adequate distances between charges. In recent years, nickel-rich cathode materials, including NCA and NCM, have emerged as candidates for LIBs due to their higher reversible capacity. Traditionally, to address the problem of capacity and performance degradation caused by electrolyte oxidation or decomposition over time and the interactions between the electrolyte and electrode during electrochemical reactions, metal oxides, such as Al₂O₃, SnO₂, ZnO, and ZrO₂, have been used as protective coatings on the surfaces of LCO, NMC, and NCA. These coatings act as a barrier to prevent harmful reactions and improve the overall stability and lifespan of LIBs [5]. These metal oxides are typically stable but electrically nonconductive. Consequently, using them as protective coatings in lithium cells increases internal resistance [6,7]. Lithium ferromanganese phosphate (LFMP), which has an olivine structure, can be used to coat nickel-rich cathode materials, such as NCA, to effectively stabilize surface stability [8] and reduce heat dissipation. LFMP also enhances corrosion and oxidation resistance to inhibit expansion and contraction during charging and discharging. However, LFMP has poor electronic conductivity. As a result, carbon-based materials that are highly conductive have been used to effectively enhance conductivity in LIBs [9,10]. Graphite variants, such as KS4 and KS6, and carbon black variants, such as acetylene black, Ketjen black, and SP, are widely regarded for their reasonable cost, excellent conductivity, and stable electrochemical characteristics. Carbon nanotubes (CNTs) have gained considerable attention in recent years due to their unique one-dimensional tubular structure and high electrical and thermal conductivities. In LIBs, CNTs are used as additives in both anode and cathode materials. CNTs have multifunctional benefits, including enhanced electron transport and increased lithium-ion insertion/removal rates through the provision of shorter diffusion pathways. By incorporating CNTs, LIBs can experience notable enhancements in their performance and efficiency [11,12]. Recent studies have demonstrated that multi-walled CNTs are effective conducting agents that can replace carbon black in high-energy LIBs. Incorporating multi-walled CNTs into NCA, NMC, and LCO electrodes has resulted in cycle life improvements ranging from 1% to 3%. This indicates that the inclusion of multi-walled CNTs can enhance the longevity and overall performance of LIBs, making them a promising choice for high-energy applications [13,14]. CNT electrodes can tolerate high currents and effectively retain capacity within the first 100 cycles [15].

In current LIB products, multi-walled CNTs are typically used at concentrations of 0.5 to 1 wt%, either replacing traditional 2 to 2.5 wt% carbon black or combined with it to enhance battery performance. Graphene has a two-dimensional structure and is being explored as a conductive additive to enhance electrode conductivity. Network structures that combine graphene and CNTs are being explored, with promising results for their application in LIBs. N-methyl-2-pyrrolidone, which is suspected of being reprotoxic, is listed in the registration, evaluation, authorization, and restriction of chemical regulation, which is a European Union regulation designed to protect human health and the environment from the risks of chemicals. N-methyl-2-pyrrolidone will be restricted in 2024 in Europe [16].

The present study employed a water-based binder system and fabricated an NCA–LFMP cathode in a lithium-ion pouch battery. Various electrochemical tests, including charge–discharge, current rate, and cycle tests, were performed on the pouch battery. Additionally, resistivity and impedance tests were performed to support the obtained results. The study investigated the effects of different amounts of conductive carbon in the electrode and achieved a high content of active material (>95%). This approach has the potential to yield a high-energy cell. The study also assessed energy density (measured in Wh/kg and Wh/L) and capacity retention.

2. Materials and Methods

2.1. Materials and Chemicals

LiNi_0.85Co_0.15Al_0.05O₂ was purchased from Ubiq Technology. CNTs were purchased from LG Chem. Styrene–butadiene rubber was purchased from Zeon. Carboxymethyl cellulose was purchased from Asland. LFMP and graphene were purchased from Sino Applied Technology. Carbon black was purchased from Imerys S.A.

2.2. Cathode Material Synthesis

An NCA slurry with a pH of less than 9 was created by mixing LiNi_0.85Co_0.15Al_0.05O₂, CNTs, styrene–butadiene rubber, carboxymethyl cellulose, and LFMP at a ratio of 94.34:0.16:2.0:2.0:1.5 (wt%) using a mixing device (Jet Paster, Nihon Spindle Manufacturing, Amagasaki, Japan). The slurry was then coated and dried on aluminum foil and reserved for subsequent use as the positive electrode (cathode). The negative electrode (anode) was created by mixing artificial graphite, styrene–butadiene rubber, and carboxymethyl cellulose at a ratio of 96:1.5:2.5 (wt%). Once dispersed using a triple-shaft planet disperser (2P-03, model 2P-1HIVIS DISPER MIX, Yumebutai, Awajishi City, Japan), the anode slurry was coated and dried on copper foil. Coating was performed using a comma coater (HIRANO TECSEED, Nara City, Japan). After being vacuum-dried overnight at 100 °C, the approximate loadings of the cathode and anode were measured at 3.1 mAh/cm² and 3.5 mAh/cm², respectively. For comparison purposes, two additional cathodes were synthesized that substituted CNTs with graphene or carbon black (maintaining the same weight ratios).

2.3. Material Characterization

X-ray diffraction was performed using a Philips PW-1700 to determine the crystalline structures of the CNTs, graphene, and carbon black-Super P cathodes. X-ray diffraction measurements were conducted at 40 kV and 40 mA with Cu Kα radiation. Raman analyses of the CNT, graphene, and carbon black cathodes were performed using a Raman Spectroscope (UniNano Tech, D2G, Yongin City, Republic of Korea). The Raman instrument operated at a laser power of 50 mW, scanning the range from 200 to 2000 cm⁻¹. Additionally, the morphologies of the graphene and carbon black cathodes were examined using a field emission scanning electron microscope (JEOL JSM-6500F, Akishima City, Japan).

2.4. Electrochemical Characterization

The freshly prepared cathode and anode materials were assembled into an LIB with a prismatic soft-packed pouch configuration. The pouch battery was approximately 3.7 mm thick, 38 mm wide, and 55 mm long (Model 404060). Nine pairs of cathode and anode electrode sheets measuring 36 × 52 mm were stacked in a zig-zag pattern with a polypropylene separator. An electrolyte was synthesized that contained 1 M LiPF₆ in EC (ethylene carbonate)/DEC (diethyl carbonate)/EMC (ethyl methyl carbonate) with a volume ratio of 2:2:1. Additionally, 3% propylene carbonate and 1% vinylene carbonate were electrolyte additives. Cells were assembled in a dry room with a dew point temperature of less than −45 °C. The assembled pouch batteries were subjected to electrochemical tests using an automatic battery cycler (NEWARE-CT-4008Q-5V6A, Milpitas, CA, USA). The tests involved discharging the cells at a constant current ranging from 200 to 3000 mA. The current rate performance tests involved charging the cells to 4.25 V using a constant current (200 mA) and constant voltage and then discharging the cells to 2.8 V. The cycling test involved charging and discharging the cells at a constant current of 1000 mA (approximately 1C of the original cell capacity). The charge and discharge cut-off voltages were set at 4.25 and 2.8 V, respectively. The open circuit voltage and alternating current internal resistance of the cells were measured using an impedance analyzer (HIOKI-3555, Nagano City, Japan) at a frequency of 1000 Hz under different states of charge.

3. Results and Discussion

X-ray diffraction profiles of the carbon black, CNT, and graphene additives are shown in Figure 1a. The profiles indicate that the additives closely resembled those of crystalline graphite in the (002) plane. However, slight shifts were observed in the peak values around 2θ = 24.8–26°. Notably, the carbon black additive exhibited a broader peak relative to the CNT and graphene additives. These peak shifts can be attributed to factors such as the oxidation treatment applied to carbon black or the curvature of the graphite plane in graphene and CNTs. Additionally, the broad peak observed at 2θ = 43.7–43.8° corresponds to the graphitic (100) crystalline lattice. Raman spectroscopy, a suitable method for investigating graphene layers in carbonaceous materials (including CNTs), was employed [17,18]. The left peaks in Figure 1b, observed at 1352 cm⁻¹ and referred to as the D band, were induced by defects and nonbasal planes in the graphitic layer. In the middle portion of the Raman spectrum, a sharp band at 1591 cm⁻¹, known as the G band, was observed. The G band is associated with in-plane symmetric carbon sp2 bonding and is typically found in highly crystalline carbon materials. In the case of the carbon black, graphene, and CNT additives, the G band peak slightly shifted to higher wavenumbers relative to normal graphite (approximately 1582 cm⁻¹), indicating a lower number of carbon plane sheet layers in the nanosized carbon additives relative to graphite. The degree of disorder in carbon materials can be estimated by evaluating the ratio of D to G band intensity. In this particular case, the ratio of D to G band intensity was higher in the CNT additive than in the carbon black and graphene additives, suggesting that different chemical vapor deposition processes are used in the production of CNTs than in those of carbon black or graphene. Additionally, a band referred to as the D′ band was observed at the shoulder of the G band. This band is a result of the pickling treatment typically applied to CNTs during their production.

The morphologies of the additives were investigated using scanning electron microscopy (SEM; Figure 2a–c). The carbon black additive had an agglomerated, irregular structure with a high degree of porosity (Figure 2a). The porosity of the material increased due to the space between the agglomerated particles. The particle size of the carbon black additive was highly variable, with an average size of 20 nm, ranging from 10 to 50 nm. The presence of such a wide range of particle sizes affects the electrical conductivity and mechanical strength of the material. The CNT additive had a highly ordered, tube-like structure with a high degree of porosity (Figure 2b). The particle size of the CNT additive was highly uniform, with an average diameter of 11 nm. This high degree of uniformity is advantageous for powering electronic devices and applications that require precise control of the material properties. The graphene additive had a highly ordered, hexagonal lattice structure with a high degree of porosity (Figure 2c). The porosity of the material was believed to be due to the presence of defects and voids in the graphene lattice. The particle size of the graphene sheets was highly uniform, with an average size of 2–8 μm.

The CNT–NCA–LFMP cathode was observed using SEM with element mapping (Figure 2d). The NCA particle size was 3–5 μm, and the LFMP particle size was 300–500 nm. LFMP particles covered the surface of the NCA, effectively stabilizing NCA surface activity and filling the voids between NCA particles. The CNTs provided a good conductive network between these cathode materials with their special linear structure. LFMP has low conductivity and requires a material with a high aspect ratio, such as CNTs, to improve its electrochemical properties. The nickel, cobalt, manganese, iron, and phosphorus elements were uniformly dispersed across the surface of the NCA (Figure 2e–i).

The electrochemical properties of the additives were investigated. Pouch batteries with a capacity of 1.1 Ah were created using NCA–LFMP electrodes and different carbon additives. The charging and discharging curves of the cells with cathodes containing 0.16% carbon black, 0.16% CNTs, and 0.16% graphene are shown in Figure 3. During the charging process, in which lithium ions intercalate into the graphite anode, the cell with the carbon black additive did not reach a discharge capacity of 1.1 Ah. Instead, it achieved a capacity of 1087.1 mAh. The cells containing CNT and graphene exhibited higher capacities than did the cells containing carbon black. The cell with CNTs achieved a capacity of 1122.9 mAh, and the cell with graphene reached a capacity of 1112.3 mAh. These results demonstrate that the incorporation of CNTs and graphene in the cathode positively influenced cell capacity, enabling the cells to surpass the 1.1 Ah threshold. Cathode unit capacity performance values, calculated by dividing the weights of the active materials by the total weights of the electrodes, are shown in

Table 1. Performance profiles of lithium-ion pouch cells with different additives.

NCA-LFMP Pouch Cells	Conductive Additive	Positive Electrode Loading (g)	NCA Active Material Weight (g)	Discharge Capacity under 200 Ma (mAh, mAh/g)	Cell Weight (g)	Cell Energy Density (Wh/kg)
NCA-LFMP/CNT	CNTs	6.217	5.931	1122.9, 189.3	18.23	224.8
NCA-LFMP/Gr	Graphene	6.227	5.941	1112.3, 187.2	18.17	223.4
NCA-LFMP/SP	Super P	6.223	5.947	1087.1, 183.1	18.28	217.1

. The CNT–NCA–LFMP cell exhibited a unit capacity of 189.3 mAh/g, and the graphene–NCA–LFMP cell had a unit capacity of 187.2 mAh/g. The carbon black–NCA–LFMP cell had a lower unit capacity of only 183.1 mAh/g. This reduced capacity may be the result of an insufficient amount of carbon black. The 200 mA discharge curve reveals that the voltage plateau was higher in the carbon black–NCA–LFMP cell (Figure 4). This suggests lower internal resistance and higher ionic conductivity in the CNT–NCA–LFMP and graphene–NCA–LFMP cells compared with the carbon black–NCA–LFMP cells. When comparing the CNT–NCA–LFMP and graphene–NCA–LFMP cells, the 200 mA discharge curves are similar, indicating comparable performance in terms of voltage behavior during discharge.

The energy densities of the cells are shown in Table 1. Despite having similar volume and weight, the CNT–NCA–LFMP cell had the highest energy density (Wh/kg). Specifically, the cell containing CNTs (0.16%) had an energy density of 224.8 Wh/kg during low-current (200 mA) discharging, approximately 3.4% higher than the cell with carbon black (0.16%). The remarkable energy density of the cell containing CNTs can be attributed to the combination of a large capacity provided by the NCA–LFMP cathode and the high voltage platform resulting from the effective conductive network built by the CNTs within the electrode. At a discharge rate of 1000 mA, the energy density of the CNT–NCA–LFMP cell still reached 220 Wh/kg. Accordingly, only 4.5–5 kg of these cells would be required to provide 1 kWh of energy.

The discharge capacities of the cells are shown in Figure 5a–d. The cell containing graphene demonstrated capacities of 1087.1 mAh and 182.2 mAh/g at 200 mA. When compared with the CNT–NCA–LFMP cell, the graphene–NCA–LFMP cell exhibited similar capacities during the 0.5C and 1C tests. However, at higher discharge rates (2C and 3C), the capacity of the graphene cell dropped significantly, with capacity retentions of only 80% and 67%, respectively. At these discharge rates (2C and 3C), the discharge capacity of the carbon black–NCA–LFMP cell was even worse than that of the graphene cell, with capacity retentions of only 77% and 46%, respectively. The 3C discharge working midpoint voltage of the CNT cell remained constant at 3.45–3.5 V (Figure 5a). By contrast, the initial working voltage of the graphene cell was approximately 3.4–3.45 V. Similarly, the carbon black cell displayed a fast decay of the working midpoint voltage, indicating a decrease in ionic conductivity, which may have contributed to the sudden capacity fading. The current in this test was 1C, representing the discharge rate for the graphene, carbon black, and CNT cells, calculated on the basis of the capacity tested under 200 mA. The CNT–NCA–LFMP cell demonstrated significantly better cycle life performance compared with the graphene or carbon black cells in this specific formulation. After 450 cycles, the CNT cell exhibited capacity retention of 91.3%, whereas the graphene and carbon black cells experienced serious capacity fading, with capacity retentions of only 90.2% and 77.4%, respectively. A possible reason for the insufficient ionic conductivity of the graphene electrode may be the inadequate length of CNTs in establishing connections between NCA–LFMP active materials. In the case of the carbon black electrodes, poor ionic conductivity resulted from both a lack of carbon additive and from the loss of particle-to-particle contact after cycling. The electrochemical properties of the NCA–LFMP cathode, when combined with carbon additives, were found to change after cycling. The lack of resilience in the sphere-like carbon black resulted in the formation of cracks inside the electrode. These cracks altered the microstructure of the electrode and led to electrical isolation between the active materials, resulting in performance degradation, including capacity and voltage fading [19,20].

In order to investigate the conductivity and the kinetic behavior of NCA-LFMP with different conductive additives, Figure 6a presents the electrochemical impedance spectroscopy (EIS) plots of NCA-LFMP-CNT, NCA-LFMP-Gr, and NCA-LFMP-SP after cycling tests. According to the equivalent circuit diagram of the inner figure in Figure 6a, the fitted data and calculated diffusion coefficients were investigated. RS is the electrolyte resistance at the highest frequency, and RSEI and RCT belong to the intermediate frequency region, corresponding to the resistance of the SEI film and the charge transfer resistance, respectively. The sloping line at low frequency represents the diffusion coefficient of the substance. According to the results shown in Figure 6a, the NCA-LFMP-SP electrode illustrated a much larger semicircle than the other samples, which indicates that a very large charge transfer resistance existed when using carbon black as a conductive additive. Both NCA-LFMP-CNT and NCA-LFMP-Gr displayed similar semicircles, which means that the electronic conductivity of NCA-LFMP-CNT and NCA-LFMP-Gr was similar. Figure 6b shows the linear relationship between Z′ and ω^−1/2 in the low-frequency region. The lithium-ion diffusion coefficient (DLi+) was calculated by the following formula:

$$D=\frac{R^2{T}^2}{2A^2{n}^2{F}^4{C}^2{\sigma }^2}$$

(1)

where R is the ideal gas constant, T is the temperature in Kelvin, F is Faraday’s constant, A represents the electrode surface, C is the concentration of Li in the electrode, and D is the diffusion coefficient. From the slope (σ), the D values for NCA-LFMP-CNT, NCA-LFMP-Gr, and NCA-LFMP-SP electrodes were calculated to be 1.06 × 10⁻¹¹, 1.8 × 10⁻¹¹, and 2.42 × 10⁻¹² cm² s⁻¹, respectively. Both CNTs and graphene exhibited better lithium ionic conductivity than carbon black.

In Figure 7a–c, the SEM images are the three electrodes after cycling tests. According to the SEM image, the NCA-LFMP-CNT electrode displayed a very compact surface on the electrode. Instead, there were many pores and cracks in the NCA-LFMP-SP electrode.

Cell safety was assessed by performing 3 mm, 8 cm/s nail penetration tests. A nail penetration test was also performed on the NCA cell without LFMP. No smoke or fire was observed during the test of the NCA–LFMP cell; however, considerable smoke and fire were observed during the test of the NCA cell (Figure 8). No damage was observed for the CNT–NCA–LFMP cell. These results indicate that the CNT–NCA–LFMP cell has excellent electrical properties and is safe. This battery is eco-friendly and has a high energy density.

Consequently, the significant improvement in the rate performance of the cells was attributed to the establishment of an effective conductive network using CNTs of sufficient length. This conductive network reduced internal resistance, facilitated ionic diffusion and transportation, and enhanced overall electrochemical reactions within the battery. The resulting improvements in electrical conductivity and diffusion coefficients led to enhanced cycling rates and increased power density, ultimately improving the cell’s performance at higher discharge rates [21,22,23] and significantly improving the rate performance of the cells.

4. Conclusions

In this work, carbon black, graphene, and CNTs were used as additives in NCA–LFMP cathodes to create NCA–LFMP–graphite lithium pouch batteries. CNTs formed a distinctive network-like structure between the active materials, which remained intact throughout charge–discharge cycling. This network structure was not observed in the electrodes that contained graphene or carbon black. Pouch cells with cathodes containing 95.4% NCA, 1.44% LFMP, and 0.16% CNTs demonstrated good tolerance to different current rates and exhibited favorable cycling performance compared with cells with the same amount of graphene or carbon black. The improved properties of the CNT cells were primarily attributed to a reduction in internal resistance, particularly the charge–transfer resistance. As the cycling tests continued, the enhanced performance could be attributed to the ability of the CNTs to stabilize the electrode structure, including facilitating interactions between the active materials. This stabilization ensured good electrical conductivity and provided an effective ionic conductive pathway, enabling sustained long cycle life under specific cycling rates. The 450 cycles of 1C testing conducted in this study meet industry standards and a reference design for commercial LIBs. Additionally, the incorporation of a small number of CNTs as a conductive additive allowed the cell to achieve higher energy density while maintaining reasonable costs in terms of additives and production processes. The CNT cells provided excellent energy density suitable for electric vehicle applications.