Single-Walled Carbon Nanotube Dispersion as Conductive Additive for Silicon-Based Lithium-Ion Batteries

Abstract

Silicon anode has recently been applied to lithium-ion batteries (LiBs) for electric vehicles (EVs), in order to improve energy density and rate capability. However, the swelling problem of the silicon anode occurs during the charging and discharging cycles. As a result, the cycle life of a battery is drastically decreased by increasing silicon loading in the anode electrode. Here we demonstrate high-quality single-walled carbon nanotube(SWCNT) dispersion as a conductive additive, in order to solve the swelling problem and thus realize long cycle life of high-energy-density battery, including silicon anode. Water-based SWCNT dispersion (0.4 wt%) is developed by mechanical separation and chemical stabilization. It forms mesh-like 3-dimension electromechanical networks over the silicon-based anode electrode. The electrode, including flexible SWCNT fibers, becomes more elastic and strong, unlike rigid multi-walled carbon nanotube (MWCNT). Therefore, the swelling problem of the silicon anode can be only solved by adding SWCNT dispersion, but not MWCNT. As a result, the cycle life of the silicon-based battery for high energy density is dramatically improved. Especially, SWCNT dispersion achieves 83% higher cycle life than MWCNT. It is revealed that the high-quality SWCNT dispersion provides great potential for high energy density and rate capability for EV batteries.

1. Introduction

1.1 Overview of SWCNT

Carbon nanotubes (CNT), a hollow nano cylinder of carbon atoms, were first reported by Sumio Iijima in 1991 [1]. Then single-walled carbon nanotubes (SWCNT), a single-layer graphene tube with a diameter of 1 ~ 2 nm, were first described by NEC [2] and IBM [3] in 1993. A CNT structure consists of sp² hybridized carbon atoms with a hexagonal lattice. CNTs are classified into single-walled CNT (SWCNT = 1 layer) and multi-walled CNT (MWCNT ≥4 layers), as shown in Figure 1.

Figure 1.

Figures of single-walled carbon nanotubes (only one layer) and multi-walled carbon nanotubes (more than four layers). The properties and characteristics of SWCNT are distinguished by MWCNT.

Chirality is defined during synthesis, and determines electrical properties of CNTs, as shown in Figure 2. Practical CNT flakes are synthesized into three types, such as metallic armchair, semiconducting zigzag, and chiral structures [4]. If the chirality is n = m, CNT becomes an armchair, which is a metallic conductor. On the other hand, when the chirality is n ≠ 0 and m = 0, CNT will be a zigzag structure, which is a semiconductor. If the chirality is random (n ≠ m and n, m ≠ 0), CNT becomes a chiral structure, which is also a semiconductor. Its band gap is inversely proportional to CNT diameter [5]. Therefore, it is essential to separate or select high-quality CNTs because the electrical conductivity of commercial CNTs differs even more than two or more times [6].

Figure 2.

The chirality of as-grown CNT fibers determines whether they are only metallic conductor (armchair) or including semiconductor (zigzag and chiral) [4].

SWCNT and MWCNT show excellent yet different properties, as shown in Table 1. SWCNT (1 layer) must be distinguished from MWCNT (≥ 4 layers) because the SWCNT shows almost better properties than MWCNT. The properties of MWCNT decrease with their structural defects as well as undefined diameters depending on the number of layers, unlike SWCNT [7].

	SWCNT	MWCNT	Comparison
Number of CNT walls [7]	1 layer	≥ 4 layers*
Density [8]	1.6 ~ 1.7 g/cm^3	1.8 ~ 2.1 g/cm^3	Al ~2.7 g/cm^3
Tensile strength [9]	~4.5 x 10^10 Pa	~4.5 x 10^10 Pa	Steel ~2 x 10^9 Pa
Electrical conductivity [10, 11]	< 10^6 S/cm	< 10^5 S/cm	Cu ~10^5 S/cm
Current density [10, 11]	> 10^9 A/cm^2	> 10^9 A/cm^2	Cu ~10^7 A/cm^2
Thermal conductivity [12, 13, 14, 15]	< 5800 W/mK	< 3000 W/mK	Cu < 385 W/mK

Table 1.

Comparison between SWCNT and MWCNT properties. Almost all properties of SWCNT are better than MWCNT as well as representative metals.

*

Properties of single-walled CNT (SWCNT = 1 layer), double-walled CNT (DWCNT = 2 layers), and triple-walled CNT (TWCNT = 3 layers) are distinguished by multi-walled CNT (MWCNT ≥4 layers).

SWCNT shape is a hollow cylinder, thus showing lower density than light metal (e.g., aluminum). Its density is determined by the number of layers and diameter of SWCNT [8]. The SWCNT structure is composed of a hexagonal lattice of sp²-hybridized carbon atoms. The covalent sp² bonds between neighboring carbon atoms make them mechanically strong. Therefore, the tensile strength of SWCNT is higher than that of steel [9]. SWCNT has an exceptionally long mean free path at room temperature of ~1 μm, thus it allows ballistic conduction [16]. Moreover, the long mean free path allows low power dissipation as well as quantum mechanical phenomena at room temperature. However, commercial SWCNTs provide different electrical conductivity because it is decreased by lattice vibration from the interlayer of MWCNT and additional scattering from impurities (e.g., amorphous carbon and aromatic hydrocarbons) arising from SWCNT synthesis. Nevertheless, SWCNT provides higher electrical conductivity than that of copper. In addition, the current density of SWCNT overcomes 10⁹ A/cm², which is bigger than that of copper due to no breakdown from electromigration [10, 11]. The strong C-C bonds and light atomic mass of SWCNT make dramatically high phonon frequencies and acoustic velocities, thus the Umklapp scattering from the phase space is unconventionally small [17]. As a result, SWCNT flows large amounts of heat by acoustic phonon modes and thus shows higher thermal conductivity than copper [12, 13, 14, 15]. Thermal conductivity in real materials is limited by the inharmonic transport of phonons. The thermal stability of SWCNT originated from a strong covalent bond between carbon atoms with a hexagonal lattice making SWCNT thermally stable. Therefore, SWCNT is thermally stable up to 400°C in the atmosphere and 1000°C in the vacuum. The thermal stability of SWCNT is enhanced with their diameter and length increase.

1.2 Synthesis of SWCNT

There are three major methods for the synthesis of CNT, such as arc discharge, laser ablation, and chemical vapor deposition (CVD) [18]. Among them, the CVD is the best method for commercialization because it is appropriate for low cost, mass production, high purity, high yield, and moderate synthesis temperature. Shigeo Maruyama has reported alcohol catalytic chemical vapor deposition (ACCVD) for the synthesis of SWCNT [19, 20]. Especially for application to a lithium-ion battery, there are key points of SWCNT flake: low price (< 1200 USD/kg), mass production (> 5 tons/year), high CNT contents (> 99%), and low metal impurities (e.g., Fe < 100 ppm).

1.3 Purification of SWCNT

During the synthesis process, CNTs are contaminated with amorphous carbon, residual metal catalysts, and fullerenes. Therefore, as grown CNT requires a purification process, such as chemical etching and physical separation [21]. The chemical purification is based on the etching of the contaminants by acids. On the other hand, the physical purification process separates contaminants from CNT by filter, for commercialization.

1.4 Dispersion of SWCNT

Dispersion means that one material (e.g., CNT) is dispersed in a continuous phase of another material (e.g., solvent, polymer, …). As grown CNT fibers tend to become aggregates and agglomerates based on the van der Walls force (Figure 3). Strongly aggregated CNT fibers require perfect dispersion in liquids for application to a commercial products and improve their performance.

Figure 3.

Conceptual figures about CNT dispersion based on distributive and dispersive mixing. SEM images show as-grown CNT aggregates and (right) highly dispersed CNT fibers.

Perfect CNT dispersion can be achieved by a combination of dispersive and distributive mixing. Figure 3 shows scanning electron microscopy (SEM) images of highly dispersed CNT fibers. As a first step of dispersion, the dispersive mixing employs mechanical energy (e.g., shear stress, impact, …) of various machines (e.g., beads mill, homogenizer, …). Therefore, the energies break and separate individual CNT fibers from as-grown CNT aggregates. After that, the distributive mixing simultaneously makes CNT fibers homogeneous spreading in liquid. At this moment, the chemical dispersion simultaneously occurs by wrapping dispersants on individual CNT surfaces. The dispersant wrapping modifies CNT surface more soluble in liquid. As a result, it enhances the dispersion quality in liquids as well as prevents reaggregation by steric hindrance or electrostatic repulsion mechanisms [22, 23]. The dispersant wrapping process is called “non-covalent functionalization” and differs from covalent functionalization accompanying defects from acid-based oxidation [24]. To commercialize the SWCNT in the battery industry, the non-covalent approach using a dispersant can realize the low cost yet high-quality dispersion. Among the three key processes (i.e., synthesis, purification, and dispersion), the dispersion process plays the most important role in the commercialization of SWCNT because performance improvement of commercial products can be achieved by making perfect dispersion of SWCNT in a liquid depending on applications.

1.5 Applications of SWCNT

SWCNT can be applied in various applications, such as electronic devices, nanocomposites, supercapacitors, and batteries [25, 26]. Among them, battery application has the largest potential market because of the big change to electrical vehicles. MWCNT and carbon black have mainly been used as a conductive additive for lithium-ion batteries. Especially, SWCNT must be distinguished from CB and MWCNT. SWCNT, a one-dimensional tube, is 1.6 nm in diameter and 5 μm in length. On the other hand, MWCNT, more than 4 tubes, shows over 5 nm in diameter and 10 μm in length. Therefore, the aspect ratio SWCNT is higher than 3000. The specific surface area of SWCNT is 1000 m²/g, which is approximately 5.5 times higher than 180 m²/g of CB [27, 28]. For this reason, SWCNT of about 9 grams can cover the surface of a football ground, as shown in Figure 4. The high specific surface area (>1000 m²/g) and high aspect ratio (>3000) of SWCNT decrease the electrical percolation threshold.

Figure 4.

Extraordinary specific surface area of SWCNT. 9.12 g of SWCNT powder covers all the surface of a football ground.

It is crucial to make high energy density and rate capability of lithium-ion battery (LiB) because electric vehicles (EV) require long mileage after fast charging. Therefore, EV batteries employ a silicon anode that has eleven times higher energy density than the current graphite anode [29]. However, silicon volume expansion (i.e., swelling problem) causes sudden degradation of cycle life of the battery [30]. Figure 5 compares the electrical network between MWCNT and SWCNT as a conductive additive of lithium-ion batteries. Although MWCNT is applied to the anode electrode, stiff anode electrode based on rigid MWCNT network can easily be broken under silicon swelling during charging and discharging (Figure 5(a)). Therefore, the electrode conductivity is decreased, and thus the cycle life of the battery is diminished. For this reason, the swelling problem arising from silicon anode has not yet been solved by MWCNT. In order to improve battery performance and thus expand the supply of eco-friendly EV, the silicon swelling problem of EV batteries should be urgently solved.

Figure 5.

Comparison between (a) MWCNT and (b) SWCNT as a conductive additive for anode electrode of lithium-ion battery. SWCNT offers higher conductivity and elastic network.

Here, we demonstrate highly dispersed SWCNT in water as a conductive additive for resolving the swelling problem. The high-quality SWCNT dispersion enables to make of elastic electromechanical networks (i.e., mesh-like) over the silicon-based anode electrode, unlike the rigid MWCNT. The mesh-like SWCNT 3D network enhances electrical conductivity as well as allows the free movement of lithium ions from silicon anodes. Moreover, it reinforces to maintain electrical network under repeated swelling and shrinking of anode electrode. Therefore, the swelling problem can be solved by making an elastic anode electrode based on SWCNT dispersion. Consequently, the cycle life of the silicon-based EV battery can be dramatically improved.

2. Dispersion of SWCNT in liquids

Single-walled carbon nanotubes (SWCNT) have some intrinsic problems of low dispersibility and agglomeration phenomenon due to strong van der Waals attraction between carbon nanotubes in their high specific surface area [31].

Although dispersing single-walled carbon nanotubes into a liquid phase is a very necessary process for many applications, the development of a practical dispersion process for SWCNT is lagging compared to other nanoparticles. Several reviews have been published on aspects of CNT solubilization and stability in solvents [32, 33]. In many studies, SWCNT dispersion is mostly achieved through the functionalization of the SWCNT surface. For example, through oxygen plasma treatment and acid treatment, SWCNT can be oxidized, and it improves its dissolution in polar solvents [34, 35, 36]. However, the industrial application of these techniques is quite problematic due to low productivity and safety concerns.

Since the main content of this text focuses on the application of SWCNT to lithium-ion batteries, this part will briefly introduce an approach to dispersing SWCNT in water, a secondary battery anode solvent.

2.1 Choosing dispersants and dosage amount

SWCNT aqueous dispersion may contain surfactants such as SDBS (sodium dodecyl benzene sulfonate) [37], CTAB (Cetyltrimethylammonium bromide) [38], and Triton X-100 or polymeric dispersants such as CMC (carboxymethyl cellulose) [39], PVP (polyvinylpyrrolidone) [40], and PAA (polyacrylic acid) [41]. They can be used alone or mixed. In non-battery fields, a dispersant with a small molecular weight is generally preferred, because it is relatively easy to remove after coating. But to apply SWCNT dispersion as a battery electrode component, a polymer-based dispersant is preferred considering long-term stability, electrode adhesion, and thermal stability.

Among various dispersants, CMC is widely chosen in the secondary battery field [42]. CMC is a very cheap and environmentally friendly biopolymer. CMC is adsorbed on the surface of SWCNT through wrapping or non-covalent functionalization and prevents agglomeration of SWCNT. The hydrophobic cellulose backbone of CMC interacts with the SWCNT surface via Vander Waals interaction, while hydrophilic sodium carboxylate moiety helps to increase solubility in water [43]. Also, it improves the dispersibility of the anode-active material in the electrode composition, preventing cluster formation of the anode-active materials, and thereby reducing electrode resistance.

Although CMC has various strengths as a SWCNT dispersant, it also has a limitation that its solution viscosity is relatively high, making it difficult to manufacture high concentrations of SWCNT dispersions. As battery cell manufacturers want to use higher concentration SWCNT dispersions to reduce costs over time, it is becoming important to find or develop dispersants that can reduce viscosity while having dispersion performance at the CMC level.

In the former paragraph, CMC was introduced as a representative dispersant for SWCNT because it is also used as a binder in secondary batteries. The dispersant must be selected according to the purpose of use, and it is very important to determine the amount of the dispersant to be used. Excessive amounts of dispersant are likely to hurt the result. Conversely, when a smaller amount of dispersant is used than the required amount, the surface of the SWCNT is not sufficiently covered and a free area is exposed, which acts as a bridging point and induces aggregation of SWCNT in the solution (Figure 6).

Figure 6.

Molecular structure of CMC.

Since various physical properties of SWCNT, such as specific surface area and aspect ratio, can affect the determination of the amount of required dispersant, the exact amount of required dispersant must be determined through multiple experiments and optimization. In an actual commercialized, for example, CMC was used at a level of 1.5 times the mass of SWCNT [44]. This dispersant dosage amount for SWCNT is generally accepted in the current secondary battery industry.

Although many dispersant manufacturers have been trying to develop dispersants for SWCNT, there are still no commercial dispersants on the market that have proven to be more effective than CMC. There may be various reasons, but the biggest reason is that SWCNT itself is very expensive and difficult to obtain supply, making it nearly impossible to secure enough SWCNT to use in developing dispersants. As the demand for SWCNT increases soon, the number of SWCNT manufacturers will also increase, and as SWCNT supply and demand become smoother, the development of dispersants for SWCNT is expected to accelerate.

Choosing the right dispersion process is as important as choosing the right dispersant. A well-dispersed SWCNT dispersion can be obtained only when the physical dispersion method and the non-covalent modification process with the help of dispersing agents are in balance.

The most representative physical dispersion processes are bead milling [45], ultrasonic cavitation [46], and high-pressure homogenization [47].

2.2 Bead milling

Equipment that uses beads as grinding media is collectively called a bead mill, and largely includes ball mills [48], attrition mills [49], and vertical/horizontal mills [50]. Bead mills are widely used in the paint, ink, chemical, pesticide, and coating industries to reduce the size of pigments and disperse them in liquids.

Bead milling can effectively promote good dispersion of CNT within the matrix, which can improve their mechanical properties and electrical conductivity [51]. Bead milling is also a scalable process, making it suitable for industrial applications. The milling effect can be increased by simply placing multiple bead mill equipment in series. The bead milling process allows for control over the dispersion quality by adjusting parameters such as milling time, bead size, and the speed of rotor rotation [52].

However, bead mills have several disadvantages. The high-energy direct impacts during the bead milling process can cause structural damage to the CNT, which can decrease their electrical properties, strengthening efficiency, and flexibility [53]. The beads used in the milling process can wear down over time, potentially contaminating the CNT dispersion and leading to inconsistent performance. After the bead milling process, the dispersion medium must be removed. This step sometimes requires additional processing, and any residual dispersion medium can affect the performance of the composite material [54].

The bead milling process has been widely used in the secondary battery industry because it is effective enough for dispersing conductive materials like MWCNT and carbon black, but it is not suitable for dispersing SWCNT. It is believed that SWCNT, which has a relatively longer fiber length than MWCNT, often blocks the screen to filter out beads in the bead mill process, and the harsh conditions of the bead mill process make the length of SWCNT too short, thereby hindering the unique advantages of SWCNT as a conductive agent (Figure 7) [56].

Figure 7.

Working principle of bead milling. Reproduced from [55] with permission from Shanghai ELE.

2.3 Ultrasonic cavitation

Ultrasonic cavitation is an often-used technique in the academic field to disperse and solubilize single-walled carbon nanotubes (SWCNT) in solution. The technique involves the use of high-frequency sound waves to create cavitation bubbles in the solvent, which then collapse and generate intense local heating and cooling cycles [57]. This process can help break down the SWCNT into smaller, more uniform sizes, which can improve their solubility and stability in solution [58]. The technique can help achieve a uniform dispersion of SWCNT in solution which can improve their stability and avoid reaggregation of SWCNT. Ultrasonic cavitation provides sufficient stress levels to separate CNT aggregates without causing many fractures in the individual nanotubes.

Although it appears to cause a small amount of damage compared to the bead mill process, it still can cause damage to SWCNT, leading to the formation of defects and cutting them into shorter pieces, which can significantly degrade their electrical and mechanical properties [59]. The technique requires a significant amount of energy to generate the high-frequency sound waves needed for the process, which can be expensive and time-consuming. Dispersing SWCNT using ultrasonic cavitation is effective only for low concentrations of dispersion. The biggest drawback of ultrasonic dispersion technology is that it is mainly used for small lab-scale experiments and may not be suitable for large-scale industrial applications [60]. Most of the well-dispersed dispersions mentioned in the previous advantages are often possible because they are manufactured on a small scale with low concentrations. When the processing amount increases or the input energy is not delivered evenly, it is common to remove large agglomerates by sedimentation through centrifugation after ultrasonic treatment [61]. This process often makes it difficult to accurately determine the final concentration of the dispersion, which is an undesirable result in practical applications (Figure 8).

Figure 8.

Working principle of ultrasonic cavitation. Reproduced from [62] with permission from the Hielscher Ultrasonics.

2.4 High-pressure homogenization

The high-pressure homogenization process is used for droplet dispersion, homogenization, emulsion, and cell wall destruction, so it is used in the electronic materials, biotechnology, pharmaceutical, food, paint, chemical, and cosmetic industries [63, 64, 65, 66]. Equipment is largely divided into a valve type and a channel type, but the basic dispersion principle is the same. In the high-pressure homogenization process, the fluid is pressurized by a high-pressure pump and then the pressurized fluid passes through a micro-orifice module or a narrow gap, where it experiences high-speed shearing, high-frequency oscillation, cavitation, and convective impact. As particles or droplets are split into nano-size and refined by these internal energies, the components of the fluid are maintained in a homogeneous and stable state of dispersion (Figure 9) [69].

Figure 9.

Working principle of (a) valve-type homogenizer [67] and (b) channel-type homogenizer [68].

The high-pressure homogenization method causes relatively little direct damage to particles, making it more advantageous for dispersing SWCNT while maintaining their physical properties compared to other processes. And its reproducibility is excellent, making it easy to scale up the process. One of the biggest advantages is that most mass-production equipment can supply fluid continuously, so there is no limit to the maximum capacity and the production speed is very fast.

However, to ensure effective dispersion with less damaged SWCNT, parameters such as treatment pressure, number of treatments, fluid feeding speed, and microchannel size must be carefully optimized. High-pressure homogenization equipment is relatively expensive compared with other dispersion equipment. Another disadvantage is that the fluid must pass through a narrow pipe and gap or microchannels, which limits the viscosity of the fluid that can be subjected to a high-pressure homogenization process [70].

There is no universal processing method for manufacturing SWCNT dispersion. Therefore, choosing the right dispersing methods and dispersant is very important and it should be based on the physical properties of SWCNT, solution viscosity, the type of matrix, etc.

As the number of companies developing and producing high-quality SWCNT continues to increase, it is expected that the application areas of SWCNT will extend beyond their purpose as conductive materials for secondary batteries. Regardless of the application field, it is highly likely that SWCNT will be used in the form of dispersion, so the functionalization of SWCNT, dispersion methods, and dispersants will be very important research areas in the future as they are now.

3. Application of SWCNT to silicon anode-based lithium-ion battery

3.1 The effect of SWCNT on the battery performance

We evaluated the performance of a silicon-based battery using conductive additives such as single-walled carbon nanotube (SWCNT), multi-walled carbon nanotube (MWCNT), and carbon black (CB).

To evaluate the battery’s performance, silicon anodes were fabricated by preparing slurries consisting of the active material (graphite: SiOx = 70:30), each conductive additive (SWCNT, MWCNT, or CB), and CMC (sodium carboxymethyl cellulose)/SBR (styrene-butadiene rubber) binder in a weight ratio of 95: x (x = the ratio of SWCNT, MWCNT, or CB): 5-x (the ratio of the CMC/SBR considering the ratio of dispersant included in dispersion solution) in DI water as the solvent. The as-prepared slurries were coated onto the Cu foil as the current collector using Mathis coater and then dried at 75°C for 12 hours. CR2032-type coin-half cells were assembled in an Ar-filled glove box with oxygen and moisture contents less than 1 ppm, using a silicon anode with a capacity of 4 mAh/cm² as the working electrode, lithium metal as the counter electrode, and separator. The electrolyte was 1 M LiPF6 dissolved in the mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) (1:1 v/v) as the electrolyte (Figure 10). The charge and discharge were performed at room temperature in the voltage range of 0.005 V (CC/CV)-1.5 V. We conducted cycling at 0.1C (3 cycles) and 0.3C (3 cycles), followed by continuous cycling at 0.5C until reaching 80% of capacity retention on the cycler test system (WONIK PNE Co., Ltd. in Republic of Korea) [71].

Figure 10.

Coin-half cell (CR2032) structure and picture of cells.

We evaluated the battery performance by incorporating 3 wt.% of carbon black (CB) into the electrode. CB is made up of spherical particles in the nanometer size range, making it easy to disperse in the electrode slurry based on water due to its low specific surface area of less than 100 m²/g. When the CB was introduced into the electrode containing 30 wt.% silicon oxide (SiOx), the initial discharge capacity was almost 100% of the designed capacity with 86% of initial efficiency. However, as the charge/discharge cycles continued from the second cycle, there was a rapid decrease in discharge capacity (Figure 11). After disassembling the cell with degraded capacity maintenance down to 90%, we observed significant structural damage to the electrode’s condition. During the washing process with dimethyl carbonate, the composite onto the current collector was easily washed away, indicating the structural instability of the silicon anode material during the charge/discharge process [72].

Figure 11.

Discharging capacities as a function of cycle number for half-cells consisting of different conductive additives, such as SWCNT, MWCNT, and CB, respectively.

The MWCNT flakes have a specific surface area of about 180 m²/g, which is approximately three times higher than that of CB. Therefore, MWCNT aggregates should be uniformly dispersed before application to anode electrode. If the dispersion is well-maintained, the required loading amount can be reduced to about one-third compared to CB. For this evaluation, MWCNT of 1 wt.% was introduced, and the results indicated that the initial discharge capacity and efficiency were similar to the result from the use of CB. However, unlike CB, there was no significant capacity drop from the second cycle, and the capacity gradually decreased over a certain period (Figure 11). The MWCNT improves the battery cycle life, which is four times higher than CB.

Furthermore, a comparative evaluation was conducted by using SWCNT as the conductive additive. The specific surface area of SWCNT is about 1200 m²/g, making dispersion extremely challenging, approximately seven times higher than that of MWCNT. However, after achieving a uniform dispersion, SWCNT of 0.1 wt.% was added, a slightly reduced amount compared to MWCNT loading. The cycle life improved by about 1.8 times compared to when MWCNT was used, based on an 80% capacity retention criterion, and about seven times compared to when CB was used (Figure 11). Moreover, unlike when CB was used, the electrode layer remained well-attached to the copper current collector, even after degrading down to 90% capacity retention. This is clear evidence that SWCNT with good elasticity made the electrode stable [73].

Although single-walled carbon nanotubes (SWCNT) have excellent conductivity and mechanical strength, it is necessary to combine spherical conductive black (CB) particles in the right proportions to create a conductive path at the contact points of active materials. To investigate the impact of combining SWCNT and Super P on battery performance, Youngseul Cho et al. formed a hybrid carbon network by mixing them in a composite electrode composed of silicon microparticles and graphite (SiMP/Gra) [74]. The weight ratio of super P to SWCNT in the electrodes was varied as 0:15, 14.5:0.5, 14:1, and 10:5, and the electrode morphology and battery performance were compared. Depending on the ratios of conductive constituents, the conductive network morphology and SiMP/Gra electrode structure were changed (Figure 12). The SiMP/Gra CNT ink 1.0 wt.% electrode showed the best rate and cycling performances with high specific capacities due to its unique hybrid carbon network (Figure 13). The addition of 1.0 wt.% of SWCNT creates an optimized combination with super P particles for the best electrochemical performance. In detail, the CNT strands tightly grasp SiMP/Gra active material and extensively connect electrode components, resulting in a durable electrical pathway throughout the electrode. By comparing a series of electrodes, this work also shows the influence of different hybrid conductive networks on the electrochemical performance and structural integrity of SiMP/Gra CNT ink electrodes.

Figure 12.

Cross-sectional SEM images of (a) SiMP/Gra + CNT ink 0.5 wt%, (b) + CNT ink 1.0 wt%, (c) + CNT ink 5.0 wt% in electrodes. SEM images of the hybrid conductive network (d) SiMP/Gra + CNT 0.5 wt% + CB(Super P) 14.5 wt%, (e) + CNT 1.0 wt% + CB 14 wt%, and (f) + CNT 5.0 wt% + CB 10 wt% in electrodes. (g) Sheet resistance analysis setup. (h) Formula of CB & SWCNT with SiMP/Gra. (i) Corresponding sheet resistance of hybrid conductive network. Reproduced from [74] with permission from the Elsevier.

Figure 13.

(a) Cyclic voltammetric peak current against the square root of scan rates. (b) Rate performance of the SiMP/ Gra-CNT ink electrodes with various current densities from 0.4 to 8 A g⁻¹. (c) Cycling performance of the SiMP/ Gra-CNT ink electrodes at 2 A g⁻¹ for 500 cycles. Reproduced from [74] with permission from the Elsevier.

3.2 The effect of dispersion technology on the battery performance

As stated earlier, SWCNT is essential in high-capacity silicon anode-based batteries. However, using SWCNT flake alone is not enough to achieve the desired effect, and it has no value as a conductive additive. To truly impart its value as a conductive additive, SWCNT must be uniformly and well-dispersed [75]. We prepared electrode slurries using three different methods: introducing SWCNT flake itself without the dispersion process, adding poorly dispersed solution, and adding fully dispersed solutions. The state of dispersion was compared based on the particle size and viscosity at SWCNT concentration of 0.4 wt.%. It was not possible to measure size and viscosity of the CNT flake because of the irregular aggregate. In case of poorly dispersed solution, size appeared to be more than 10 um and viscosity was above 2000 cP. While, the size was about 3 ~ 5 um, and the viscosity was around 250 cP when the SWCNT was well-dispersed under our condition, which means the solution shows flowability, as shown in Figure 14.

Figure 14.

Digital image of well-dispersed aqueous SWCNT solution showing flowability.

We then observed the electrode’s morphology and examined the impact of dispersion on battery performance through battery charge/discharge cycle life tests (Figure 15). The results showed that when we added SWCNT flake itself, the slurry mixing process alone was insufficient to disperse the agglomerated SWCNT flakes. Because of inadequate wetting, coating the electrode was impossible (Figure 15(a)). When we added the partially dispersed dispersion solution, agglomerated particles were still present, causing surface scratching problems while coating the electrode (Figure 15(b)). Nevertheless, we chose areas with good coating conditions for coin-cell fabrication and performance evaluation. The results were similar to those of well-dispersed MWCNT 1.0 wt.% in terms of charge/discharge cycle life (Figure 15(d)). Introducing a uniformly well-dispersed dispersion solution allowed us to manufacture electrode slurries without any problems and resulted in excellent battery performance (Figure 15(c,e)). This highlights the importance of complete dispersion while using SWCNT as a conductive additive in silicon-based electrodes.

Figure 15.

Digital images of slurries containing (a) SWCNT flake itself, (b) poorly dispersed solution, and (c) well-dispersed solution, respectively, coated onto Cu current collectors. Cycling performance of the cell based on (d) poorly dispersed solution and (d) well-dispersed solution at 0.5C.

The development of dispersion technology is crucial to effectively disperse single-walled carbon nanotubes (SWCNT). We experimented to examine the impact of applying different dispersant formulations to the same SWCNT flakes on battery performance (Figure 16). We evaluated the cycling performance of a coin-half cell based on two dispersion solutions; one with CMC (sodium carboxymethylcellulose) and the other with our patented formulation applied to SWCNT flakes as a conductive additive.

Figure 16.

Cycling performance of coin-half cell made from dispersion solution consisting of (a) CMC and (b) our patented formulation.

We compared the viscosities of the dispersion solutions at the same SWCNT concentration (0.4 wt.%). The viscosity of the dispersion solution with dispersant CMC was 1393 cP, while the viscosity of the dispersion solution with our patented formulation decreased to 76 cP. This indicated that our patented formulation had better wettability for SWCNT surface than CMC. This increased the solid content of the electrode slurry by about 2–3 wt.%, which could potentially improve the drying speed of the electrode by approximately 10%.

We evaluated battery performance using each dispersion solution and found that the initial capacity and efficiency were similar (about 660 mAh/g and 86%, respectively). However, the cell consisting of an electrode with our patented formulation showed an improved charge/discharge cycle life from 42 to 52 cycles. This enhancement in cycle life suggests a synergistic effect where the superior mechanical properties of the dispersant inhibit silicon volume expansion, maintain adhesion of the anode material elastically, and create a more evenly distributed three-dimensional conductive network due to the better dispersibility induced by the dispersant. Therefore, it can be concluded that the development of dispersants is a key technology when dispersing SWCNT in water and applying it to silicon-based batteries. Additionally, excessive energy during dispersion should not cause severe damage to the SWCNT surface or morphology. While low viscosity may indicate effective dispersion, damage to the SWCNT surface can hinder its original electrical conductivity and mechanical properties, resulting in marginal improvements.

4. Conclusion

In this chapter, we have covered essential aspects: (1) methodology of SWCNT dispersion in liquid and (2) application of SWCNT dispersion to silicon-based lithium-ion battery (LiB) for electric vehicle (EV).

High-quality dispersion of SWCNT is highly complicated because as-grown SWCNT fibers are strongly aggregated. SWCNT aggregates are mechanically separated by supplying high energy (i.e., shear and impact). At that moment, the surface of separated SWCNT fibers in water is simultaneously stabilized by wrapping polymeric dispersants (i.e., CMC and self-developed dispersant). Consequently, we have developed 0.4 wt% SWCNT dispersion in water. The high-quality SWCNT dispersion is applied to a silicon-based battery as a conductive additive. It forms a 3-dimension electromechanical network over the anode electrode and thus the electrode became more elastic and strong, unlike the network, including rigid MWCNTs. As a result, cycle life degradation problem is solved by suppressing the swelling problem of silicon anode under repeated charging and discharging. Moreover, the SWCNT dispersion dramatically enhances the battery cycle life, which is 83% higher than MWCNT.

It is verified that the highly dispersed SWCNT offers excellent potential to realize high energy density and rate capability of LiB for long-distance mileage (> 640 km) and fast charge (< 15 min) of EV. The SWCNT dispersion can be used for improving the electrical and thermal conductivity of not only anodes but also cathodes, separators, modules, and packs of batteries. It can also be used in different application areas, such as coating materials of cathode for dry process and conductive additives for next-generation batteries (e.g., all-solid-state battery). Almost all industries should consider exploiting the excellent properties of SWCNT for their performance improvements. Critical obstacles to the commercialization of SWCNT have not yet been overcome, such as cost effective synthesis and high quality dispersion of SWCNT. Nevertheless, we need to pay attention to SWCNT because it is only a question of time before the commercialization is performed.