Selection of Optimum Binder for Silicon Powder Anode in Lithium-Ion Batteries Based on the Impact of Its Molecular Structure on Charge–Discharge Behavior

Abstract

The high-capacity and optimal cycle characteristics of the silicon powder anode render it essential in lithium-ion batteries. The authors attempted to obtain a composite material by coating individual silicon particles of µm-order diameter with conductive carbon additive and resin to serve as a binder of an anode in a lithium-ion battery and thus improve its charge–discharge characteristics. Structural strain and hardness due to stress on the binder resin were alleviated by the adhesion between silicon or copper foil as a collector and the binder resin, preventing the systematic deterioration of the anode composite matrix immersed in electrolyte compositions including Li salt and fluoride. Moreover, the binder resin itself was confirmed to play a role of active material with occlusion and release of Li-ion. Furthermore, charge–discharge characteristics of the silicon powder anode active material strongly depend on the type of binder resin used; therefore, the binder resin used as composite material in rechargeable batteries should be carefully selected. Some resins for binding silicon particles were investigated for their mechanical and electrochemical properties, and a carbonized polyimide obtained a good charge–discharge cyclic property in a lithium-ion battery.

1. Introduction

Secondary cells such as lithium-ion and fuel cells are widely used as rechargeable batteries for home appliances [1,2,3,4,5,6], automobiles [7,8,9,10,11,12,13,14], smart houses [15,16,17,18], etc. These batteries can be made compact and highly efficient by designing them using high-capacity electrical energy and uniformly thin electrode (cathode and anode) films with a large area. Research focusing on the synthesis and selection of the active materials and conductive additives used in these batteries is being promoted. However, when using actual batteries (e.g., in an experiment), establishing a stable assembly technology for electrode composite materials using a wet process is necessary. In terms of compatibility between the active materials, conductive additives, and binders, the selection varies based on the combination, composition ratio, and application (batteries consisting of a solid [19,20,21,22,23,24,25,26], polymer [27,28,29,30,31,32], liquid electrolyte [25,31,32,33,34,35,36,37], etc.). However, there is no specific analysis of the impact of the binders’ molecular structure and the combination of anode active materials on the charge–discharge behavior of rechargeable batteries. Thus, we used a binder resin, a matrix material supporting electrode active materials, and attempted to select an optimum structure for silicon (Si) powder anodes [38,39,40,41,42,43] to improve the electrochemical properties of lithium-ion rechargeable batteries.

It is assumed that the ideal structure for this binder, as a matrix that supports active materials and conductive additives, is a polymer that constitutes the binder spreading as a net and supporting active materials and conductive additives. The Si anode undergoes repeated and rapid expansion and contraction due to the storage and release of lithium ions [44,45,46,47], thus separating from the binder. As a result, the binder is no longer able to support active materials and the electrode’s structure collapses. This leads to the rapid deterioration of the charge–discharge behavior. Therefore, we chose various polymers based on commercially available Si powder (several micrometers in diameter), which are typically employed as binders in lithium-ion rechargeable batteries (Kojundo Chemical Laboratory Co., Ltd., Saitama, Japan), as anode active materials. We examined the impact of their molecular structures on the charge–discharge behavior of Si anodes.

With regard to anode active materials that use Si powder substrates, we propose herein an ideal electrode matrix structure that can improve the charge–discharge behavior of Si. Using multifaceted evaluation methods, we searched for a structure of binder resin that showed positive impacts on the charge–discharge behavior of Si powder anode active material.

Si, an active material of a rechargeable battery anode, shows charge–discharge behavior with a high capacity of approximately 4000 mAh/g [38], and is garnering attention as an active material for lithium-ion rechargeable battery anodes instead of graphite. Si, as an active material, repeatedly stores and releases lithium ions, but when lithium ions are incorporated into Si, the volume of the active material drastically expands [48] causing the rapid deterioration of the charge–discharge behavior. Presently, studies on suppressing such deterioration are being conducted [49], and charge–discharge behavior is being improved. However, the matrix structure that supports Si active material within electrodes is destroyed by volume expansion and contraction of Si, and thus it is difficult to support the electrode structure over a long period for practical use. Therefore, with a Si powder anode as the active material, a matrix structure is needed where electrode films can alleviate the stress even if there is expansion and contraction of Si. Figure 1 shows an ideal matrix structure that supports conductivity and ion conductivity, while being a buffer material for a volume change in Si active material, as a polymer matrix net. To build this structure, the selected binder should have a low electrochemical loss due to lithium consumption and a suppressed expansion of composite material due to expansion and contraction of active material. We therefore performed a thorough examination to select an optimum binder in terms of adhesion, extension, compression, and electrochemical properties.

For our study, we chose some resins (binders) of polyvinylidene difluoride (PVDF), polyamide-imide (PAI), polyimide (PI), thermoplastic polyimide (TPI), polybenzimidazole (PBI), polyacrylonitrile (PAN), and polyacrylic acid (PAA) that are used as wet films for battery electrodes. Using the materials mentioned above as binders, we attempted to evaluate the mechanical and electrochemical properties of resins.

2. Experiments

2.1. Shear Peeling

We mixed and stirred Si powder (3N powder approximately 5 μm; Kojundo Chemical Laboratory Co., Ltd., Saitama, Japan) and a conductive additive (Ketjenblack® (KB) EJ300J; Lion Specialty Chemicals Co., Ltd. Tokyo, Japan) (weight ratio of Si:binder:KB, 1–25:1.5:1), and created a sample in Figure 2 by sandwiching it between Cu foils of 80 µm thickness (The Nilaco Corporation, Tokyo, Japan) for the shear peeling test. After sintering, the thickness of the binder film with Si powder and KB was 120 µm, and the sample was immersed in 40 °C electrolytes (1 mol/L lithiumhexafluorophosphate (LiPF₆), ethylene carbonate (EC): dimethyl carbonate (DMC): fluoroethylene carbonate (FEC): difluoroethylene carbonate (DFEC), 2:5:2:1 v/v%) for 3 h, and shear peeling strength was evaluated using a tensile test (load cell 5848; Instron Corporation, Norwood, Massachusetts, USA). As a result, we confirmed that, regardless of the electrolyte material compositions, adhesion improved with the added Si. In addition, we measured dependence of peel strength on electrolyte compositions (Li salt/fluoride) with 3.0 g of Si.

Figure 3 shows the result of the shear peeling test. Polyvinylidene difluoride (PVDF) and polybenzimidazole (PBI) were peeled before being immersed in the electrolyte. There were notable differences for the other materials based on electrolyte immersion time, the amount of Si, and the presence/absence of Li salt in the electrolyte. Although the presence/absence of Li salt in the electrolyte did not have much impact on adhesion, the amount of Si in contact with the Cu foil and binder and the fluorine exposure time are considered important parameters.

2.2. Compressive Strength

We prepared a pellet-type composite material with a 7 mm diameter and a 3 mm thickness with an electrode-like composition, and verified the composite strength of Si–resin adhesion strength and strength of resin alone through compression testing. Figure 4 shows the correlation between modulus of elasticity and strain rate of pellets due to compression before immersion and after 60 min of immersion in 40 °C electrolyte.

Figure 5 shows the result of the above-described compression test. The pellet-type polyamide-imide (PAI), polyacrylonitrile (PAN), and polyacrylic acid (PAA) show attenuation of compression resistance due to electrolyte immersion. In addition, similar to the shear peeling test (Figure 2), compression resistance improved with the added amount of Si.

We further prepared a resin film (100 mm², 10 m thick), and soaked it in the electrolyte including EC, DMC, FEC, and DFEC at 40 °C for 2 h. Subsequently, we evaluated the rate of weight increase after soaking (Figure 6). It showed that materials with simpler molecular structure had a higher expansion rate.

2.3. Evaluation of Electrochemical Properties of Resins as Active Materials

We examined the dependence of charge–discharge properties on the binder resins. We evaluated the charge–discharge properties and cycle properties of retention, shown in Figure 7. For the anode, the mixing weight ratio of binder:conductive additive (copper powder; Kojundo Chemical Laboratory Co., Ltd., Saitama, Japan) was 1:8. Electrochemical testing of the composite anode electrodes was conducted using two-electrode 2032 type test coin cells (Housen Corporation, Osaka, Japan), caulking metal cups with a gasket to maintain the electrode assemblies, a polypropylene/polyethylene stacked film separator and a 1 M lithiumhexafluorophosphate (LiPF₆) electrolyte within solvents of DFEC/FEC/EC/DMC = 1/2/2/5 v/v%. The electrochemical performance of the two-electrode coin cells with LiCoO₂ cathode was evaluated using a potentiostat machine (Hokuto Denko Co. Ltd., Tokyo, Japan). The initial electrochemical performance of the composite anode in Figure 7a was evaluated in a coin cell using a constant current charge–discharge test in the voltage range of 0.3–2.0 V with a current density of 0.075 mA/cm² at room temperature. The state of the charge was set to 100%. The charge–discharge property of Figure 7a indicates the measured result using PI in an anode. We observed a tendency of the properties to change due to the binder resin in Figure 7b.

Each binder has a function of active material for the lithium-ion battery, and the irreversible capacity was large. In particular, binders, which had no benzene ring in the binder’s molecular skeleton, had low retention, as shown in Figure 7b. It is presumed that a binder’s benzene ring molecular structure had a role of occlusion and release of lithium ion in the graphite-like structure as anode active material.

Figure 8 shows the evaluation of the dependence of cycle properties on the composition of binder resins and the amount of conductive additives. A 2032 type coin cell using LiCoO₂ as a cathode electrode and a 1 M lithiumhexafluorophosphate (LiPF₆) electrolyte was used to evaluate the cyclic properties of charge–discharge. For the anode, the mixing ratio of Si power:binder:conductive additive (KB) was 10:1.5:1. The electrochemical performance of the composite anode shown in Figure 8 was evaluated in a coin cell using a constant current charge–discharge test in the voltage range of 1.6–4.1 V with a current density of 1.5 mA/cm² at room temperature. The state of the charge was set to 60%. We observed a tendency of the properties to change due to the binder resin in an anode electrode.

The cross-sectional scanning electron microscopy (SEM) views with overviews of the anode electrode as insets in each SEM image of Figure 9 show the status of the anode electrodes after the 100th cyclic charge–discharge measurement. Figure 9a showing PVDF as a binder had an unclear boundary between Si powders and PVDF by the crystal collapse owing to the expansion and contraction of Si in the composite. It seems that the shape of Si powders in polyimide (PI) shown in Figure 9b was maintained; however, some small cracks on the surface of each Si powder occurred. It is believed that the binder layer covering each Si particle plays a buffering role by preventing cracking following drastic changes in volume with numerous repetitions of the charge–discharge cycle. We found that the polymer having large molecular structures and strong adhesion was effective in preventing the crystal collapse of Si by occlusion and exudation of Li-ion.

3. Results and Discussion

Figure 10 shows the correlation between the mechanical and electrochemical properties of the binder sample and the reversible capacity of the composite with binders, Si powder, and a conductive additive at the 100th cycle, as discussed and shown in Figure 8. Each sample was constructed (a) by sandwiching the composite of binders, Si powder, and a conductive additive between Cu foils for shear peeling strength, (b) using the pellet-type composite material including binders, Si powder and a conductive additive for compressive strength, and (c) using the composite with binders and copper powder for the retention at 10th cycle.

The horizontal axis denotes the reversible capacity of Si powder, binder, and KB at the 100th cycle, and the vertical axis denotes (a) shear peeling strength and (b) compressive strength of each resin, and (c) retention between charge and reversible discharge of binders. It shows that cycle characteristics strongly depend on each of the mechanical and electrochemical measured results. This indicates that when the volume of the Si active material rapidly expands or contracts, the structural strain caused by the stress on the binder resin is attenuated due to the adhesiveness between Si and the binder resin and the strength of the resin, preventing the systematic destruction of the matrix.

In the case of Si, the irreversible capacity during initial charge–discharge is about 20% of the total capacity [50]. To reduce the irreversible capacity of an electrode of a rechargeable battery, selecting a binder resin with low irreversible capacity is an important factor when designing an electrode.

When the electrode expands due to expansion in the Si volume while charging, the overall matrix, including the binder resin and conductive additive, also expands. Although the volumetric energy density changes as the electrode expands, there is no obvious correlation with the composition of the binder resin (i.e., resin strength).

Based on the molecular structures of each binder resin and the above-described experiment, we discovered that the type of side chain in PVDF, PAN, and PAA has a strong impact on material strength. In the case of PVDF, immersion in electrolytes containing fluorine leads to incompatibility with Si and rapid changes in the volume of the matrix that makes up the electrode. PAI, thermoplastic polyimide (TPI), and PBI have large structures but the main chain is asymmetrical, as shown in Figure 11, where solvents can easily enter the gap. These resins contain a six-membered ring and have high molecular weight, justifying their high hardness. Moreover, the asymmetrical structure leads to expansion in the electrolyte, which results in swelling as the electrolyte infiltrates. Furthermore, when hydrogen is present in the side chain, only relatively weak bonds, such as hydrogen bonds, exist with the dangling bond on the surface of Si, representing Van der Waals forces. In such cases, roughness on the surface of Si provides an anchoring effect in bonding, making it prone to peeling due to expansion and contraction of the substrate (Si) and insertion of fluorine between layers. On the other hand, if there is oxygen in the side chain, a relatively strong bond is supported by covalent bonding in Si. Thus, even when immersed in the electrolyte, adhesion with Si can be maintained, as shown in Figure 12.

Table 1. Comparison with electrical and ionic conductance between each binder.

	PVDF, PAN, PAA	PI, PAI, TPI, PB
Electrical Conductance	Small	Large
Ionic Conductance	Small	Large

shows the comparison with electrical and ionic conductance between each binder. It is assumed that the electrical and ionic conductance between the active materials supported by the binder resin is regulated by the resin material. The specification of binders existing between the active materials impacts the electrical and ionic conductance between the active materials as an electrode, leading to an impact on cycle characteristics, as shown in Figure 8.

Based on the above evaluations, we successfully examined the dependence of charge–discharge behavior of active materials on binder resins. Through our study, we found that to obtain load characteristics that are relatively stable in a mechanical and electrochemical condition, the proper selection of molecular structure and composition for a matrix that supports Si active material is necessary. In short, the appropriate selection of molecular structure improves the charge–discharge behavior, presenting an idea for selecting binder resin for other active materials, in addition to Si.

4. Conclusions

In our study, we evaluated the mechanical and electrochemical behavior of binder resins used in anode active materials for rechargeable batteries. Specifically, adhesion of the binder resin and strength of the resin are responsible for improving the charge–discharge cycle characteristics of Si anode active material. This indicates that, during rapid volume expansion and contraction of Si active material, structural strain due to stress on the binder resin is alleviated by the adhesion between Si and the binder resin and hardness of the resin, preventing the systematic deterioration of the matrix. Furthermore, charge–discharge characteristics of Si anode active material strongly depend on the type of binder resin used; therefore, the binder resin used as composite material in rechargeable batteries should be carefully selected.