Research on improving the cycle performance of silicon-based negative electrodes for lithium-ion batteries

1. Introduction

Silicon has the highest theoretical specific capacity (4200 mAh g-1) and the lowest delithiation potential (<0.5 V), making it one of the most promising negative electrode materials for lithium-ion batteries to replace graphite. [1] However, during the charge and discharge process, silicon undergoes a huge volume change, resulting in material pulverization, peeling, loss of electrical contact, and rapid capacity decay. [2] In order to reduce the volume effect of silicon materials, people have tried a variety of methods, including the preparation of amorphous silicon films, nano-silicon, porous silicon, silicon oxides, silicon-containing non-metallic compounds, silicon-containing metal compounds, silicon/carbon composites, silicon/metal (active or inert) composites, etc. [3, 4] These methods either inhibit the volume expansion of silicon materials or improve the electrical contact between silicon particles, thereby improving the cycle stability and initial charge and discharge efficiency of silicon negative electrodes to a certain extent. In addition to the improvement of silicon-based active materials themselves, people have also done a lot of research on electrode preparation processes and electrolytes.

1.1 Improvement of electrode preparation process

Generally, commercial electrodes are composed of active materials, conductive agents and binders. The conductive agents are dispersed in the binders to make them electronically conductive, while the binders tightly wrap the active material particles to prevent them from pulverizing and loosening during the cycle and losing electrical contact. The cycle stability of the electrode is not only related to the active material, but also greatly affected by the properties (strength, elasticity, adhesion, etc.) and distribution state of the binder.

In addition to the volume effect of the silicon material itself and the electrical contact state between the internal particles, the contact state between the silicon material and the current collector also has a great influence on the cycle stability of the negative electrode. There are two main ways to improve the electrical contact state between the silicon material and the current collector:

One is to improve the performance of the binder. Traditional polyvinylidene fluoride (PVDF) binders easily absorb electrolytes and swell, resulting in decreased bonding performance. On the one hand, new high-performance binders are explored. For example, 1% styrene butadiene rubber (SBR)/1% sodium carboxymethyl cellulose (SCMC) is used as a binder to prepare carbon-coated silicon anodes. Its cycle performance is better than that of electrodes using 10% PVDF binder. This is because SBR+SCMC has greater elongation and adhesion. [5] On the other hand, the study modifies the traditional PVDF binder to improve its strength, elasticity and viscosity, which plays a role in inhibiting the volume change of the active material, enhancing the bonding force between silicon particles and the current collector, and improving electrical contact. The modification methods mainly include cross-linking and heat treatment. [6] The charge and discharge cycle performance of silicon anodes can be improved by improving the performance of the binder, but this improvement is very limited and far from the requirements of commercialization.

The second is to change the surface morphology of the current collector. The greater the surface roughness of the current collector, the greater the conductive contact area between the active material and the current collector, the higher the adhesion strength, the less likely it is to peel off during the charge and discharge process, and thus has higher cycle stability. Kim YL [7] used the same silicon-carbon composite material and binder (PVDF) to prepare silicon-based negative electrodes on current collectors with different surface morphologies, and their initial capacities were all around 800 mAh/g. After 30 charge and discharge cycles, the reversible capacity of electrode a with a flat current collector has decayed to 300 mAh/g, the capacity of electrode b with a rough current collector has decayed to 650 mAh/g, and the reversible capacity of electrode c with a nodular protrusion on the current collector surface remains at 800 mAh/g. However, the preparation of this copper current collector with a nodular surface requires two electrodepositions, which is a more complicated process and increases production costs.

This paper designs a new type of silicon-based negative electrode structure, using flexible acetylene black coating instead of traditional copper foil as the current collector, bonding the active material between the acetylene black coating and the polyethylene film to improve the bonding strength between the active material and the current collector, and hopes that this sandwich structure can buffer the volume change of the silicon-based negative electrode during the charging and discharging process.

1.2 Research on electrolyte film-forming additives

Another important reason for the capacity decay of silicon-based negative electrodes is that the decomposition of LiPF6 in the electrolyte produces trace amounts of HF that corrodes silicon, see equations (1) and (2). [8]

(1)

(2)

In addition, since silicon has difficulty forming a stable surface solid electrolyte (SEI) film in conventional LiPF6 electrolyte, along with the destruction of the electrode structure, a new SEI film is continuously formed on the exposed silicon surface, exacerbating the corrosion and capacity decay of silicon. [9]

The formation of SEI film on the electrode surface is due to the electrochemical reduction of organic salts and solvents. Its morphology and composition are mainly determined by the components in the electrolyte. It not only affects the dynamics of lithium insertion and extraction of the electrode, but also affects the surface stability during long-term cycling. [10, 11] However, there are still few studies on SEI film on silicon surface and film-forming additives. [12, 13] This paper uses silicon thin film as electrode to study the effects of four electrolyte additives on battery cycle performance, and makes a preliminary exploration to improve the interfacial compatibility between silicon and electrolyte.

2. Experiment

2.1 Preparation of silicon-based anode with flexible current collector

0.85g of carbon-coated silicon was mixed with 0.05g of conductive carbon black and 10g of styrene-butadiene rubber-sodium carboxymethyl cellulose emulsion (solid content 1wt%) for 6 hours, then coated on the surface of a 20-micron-thick porous polyethylene film, and vacuum dried at 40°C for 8 hours to obtain a carbon-coated silicon layer. 0.9g of acetylene black and 10g of styrene-butadiene rubber-sodium carboxymethyl cellulose emulsion (solid content 1wt%) were mixed and stirred for 4 hours, then coated on the surface of the carbon-coated silicon layer, and vacuum dried at 50°C for 10 hours to obtain a silicon-based negative electrode with a flexible current collector (i.e., acetylene black coating).

Figure 1 Schematic diagram of the silicon-based negative electrode structure with a flexible current collector

The conventional negative electrode preparation method with copper foil as the current collector is as follows: 0.85 g of carbon-coated silicon is mixed with 0.05 g of conductive carbon black and 10 g of styrene-butadiene rubber-sodium carboxymethyl cellulose emulsion (solid content 1 wt%) and stirred for 6 hours, then coated on the surface of 20-micron thick copper foil, and vacuum dried at 80°C for 10 hours to obtain a silicon-based negative electrode with copper foil as the current collector.

2.2 Preparation of thin film silicon anode

The thin film silicon negative electrode was provided by the Laser Chemistry Institute of Fudan University. Using copper foil as the substrate (the surface was previously etched with 0.12 mol L-1 FeCl3 solution) and silicon wafer as the target, an amorphous silicon film was obtained by plasma sputtering in a vacuum environment.

2.3 Preparation of electrolyte

The conventional electrolyte was purchased from Zhangjiagang Guotai Huarong Co., Ltd. and was a mixed solution of 1 mol·L-1 lithium hexafluorophosphate in ethylene carbonate and dimethyl carbonate (volume ratio 1:1). Four additives were added respectively, and their names and amounts are shown in Table 1.

Table 1 Full name, molecular formula and amount of additives used

2.4 Battery Assembly and Testing

The silicon negative electrode with a flexible current collector was punched into a pole piece with a punch of 12 mm in diameter, placed in a vacuum oven and dried at 40°C for 8 hours, and then transferred to a glove box filled with argon. The metal lithium sheet was used as the counter electrode, ET20-26 was used as the separator, and a conventional electrolyte was used to assemble a CR2016 button cell. The constant current charge and discharge performance test was carried out on a LAND battery test system (provided by Wuhan Landian Electronics Co., Ltd.) at 20°C. The charge and discharge cut-off voltage was 0.01~1.5V relative to Li/Li+, and the charge and discharge current density was 0.2 mA cm-2.

For the thin-film silicon anode, a half-cell was assembled using an electrolyte containing additives, with a charge and discharge current density of 0.0381 mA cm-2 and a voltage range of 0-1.5 V. Other conditions are the same as above. In this study, discharge corresponds to the lithium insertion process of the material, and charging corresponds to the lithium removal process.

2.5 Observation of electrode surface morphology

The surface morphology of the silicon thin film electrode after 20 cycles was observed using a field emission scanning electron microscope (JSM-7401F). The electrode in a completely delithiated state was taken out of the battery in a glove box, repeatedly cleaned with DMC, dried naturally, placed in a sealed glass bottle filled with argon, then taken out of the glove box and quickly transferred to the sample chamber of the scanning electron microscope.

3. Results and Discussion

3.1 Silicon-based anode with flexible current collector

The charge and discharge capacity of the silicon-based negative electrode with acetylene black coating as flexible current collector in the first 20 cycles is shown in Figure 2. Its first charge and discharge coulomb efficiency is 85%, and the reversible capacity after 20 cycles is about 1100 mAh g-1. The silicon-based negative electrode with conventional copper foil as current collector has a first charge and discharge coulomb efficiency of 81%, and the reversible capacity decays to less than 200 mAh g-1 after 12 cycles.

The improvement of cycle performance is due to the sandwich structure of this silicon-based negative electrode. The active material carbon-coated silicon layer is tightly sandwiched between the elastic separator and the flexible current collector layer. The flexible current collector layer changes accordingly with the deformation of the carbon-coated silicon layer, reducing the mechanical stress between the two, thereby improving the interface electrical contact state and significantly improving the cycle stability of the silicon-based negative electrode.

Figure 2 Cycling performance of silicon-based anodes using flexible current collectors and conventional current collectors

3.2 Electrolyte film-forming additives

The first charge and discharge curves of batteries containing different additives are shown in Figure 3, and the data such as the lithium storage capacity and the first charge and discharge efficiency are detailed in Table 2. Each sample has an obvious platform at around 0.5V, corresponding to the formation of the SEI film on the surface of the silicon film. Below 0.5V, the lithium insertion process of amorphous silicon is mainly carried out. The first charge and discharge efficiency of the battery without additives is 60.6%. After adding VC and ET, the first efficiency is increased to 66.6% and 61.2%; after adding SO2Cl2 and LiBOB, the first efficiency drops to 51.7% and 49.2%, respectively. It can be seen that the irreversible capacity consumed when forming the SEI film after adding different additives is different.

Figure 3 First charge and discharge curves of silicon thin film batteries containing different additives

Table 2 Comparison of electrochemical performance of silicon thin film batteries containing different additives

The delithiation capacity-cycle number curves of each sample are shown in Figure 4. After adding VC, SO2Cl2 and LiBOB to the conventional electrolyte, the delithiation capacity retention rate of the silicon thin film battery after 100 cycles increased from 37.4% to 83.3%, 51.2% and 44.4%, respectively, but the addition of LiBOB and SO2Cl2 reduced the battery capacity. The addition of ET did not improve the cycle performance.

Figure 4 Cycling performance of silicon thin film batteries containing different additives

In order to find out why different additives affect the performance of silicon thin film batteries, we conducted a scanning electron microscope observation on the silicon film after 20 cycles, as shown in Figure 5. Figure 5 (a) shows the surface of the silicon film after adding VC. The surface is relatively dense but rough. Further magnification also shows that the surface is covered with tumor-like protrusions. This may be due to the aggregation of the SEI film formed by the VC-containing electrolyte during the growth process. After adding VC, the composition of the SEI film changes, and polyalkyl lithium carbonate polymers are generated, making the SEI film more flexible and adhesive. [14] The SEI film formed in conventional electrolytes is mainly composed of inorganic lithium salts and a small amount of alkyl lithium, and has poor flexibility. In contrast, the SEI film formed by VC-based electrolytes is more able to adapt to the volume change of silicon during the charge and discharge process without breaking, thereby improving the cycle stability of the battery.

Figure 5 (b) shows the surface of the silicon film after adding SO2Cl2. It can be seen that there are many holes on the electrode surface, which may be due to the decomposition of SO2Cl2 at a higher potential to release SO2 gas, see formula (3). The cycle performance of silicon thin film batteries containing SO2Cl2 is still poor, which may be because the SEI film formed by SO2Cl2 is loose and porous, and the mechanical strength is not high enough, so it is easy to break and fall off.

(3)

Figure 5 Surface morphology of silicon film after 20 cycles of batteries containing different additives (the inset is an enlarged image)

(a)VC; (b) SO2Cl2; (c) LiBOB; (d) ET

The surface morphology of the silicon film after adding LiBOB and ET is shown in Figure 5 (c) and (d), respectively. From the magnified electron microscope image, it can be seen that the surface of the sample containing LiBOB is relatively flat, while the surface of the sample with ET has traces of damage and erosion. This may be because the electrolyte containing ET is difficult to form a complete SEI film on the silicon surface, so that the exposed silicon is corroded by the trace amount of HF in the electrolyte, resulting in rapid capacity decay. LiBOB can form a relatively uniform SEI film on the electrode surface during the reduction and decomposition process in an organic solvent, as shown in Figure 5 (c). However, the battery capacity decay of the battery with LiBOB added to the electrolyte is still relatively fast.

In order to find out why the SEI film was formed but the cycle performance of the silicon film was not improved, we analyzed the charge and discharge efficiency of batteries containing different additives in the first 20 cycles, as shown in Figure 6. The battery containing VC electrolyte has the highest charge and discharge efficiency (about 99%), followed by the battery without additives and the samples containing LiBOB, ET and SO2Cl2. In the first 10 cycles, the charge and discharge efficiency of the battery containing LiBOB was significantly lower than that of the battery containing VC. It can be seen that the battery containing LiBOB not only consumes a lot of irreversible capacity when the SEI film is formed in the first cycle (see Table 2), but also loses a lot of capacity in each cycle. This may be because the SEI film formed by LiBOB has low mechanical strength and breaks during the volume change of silicon, and it needs to continuously form an SEI film for repair. This will increase the thickness of the SEI film on the silicon surface, but the effect of improving the cycle stability of silicon is limited. The samples with ET and SO2Cl2 have lower charge and discharge efficiencies than those without any additives. It may be that they have more side reactions and cannot form a dense, stable and reliable SEI film, so they do not play a good protective role for the silicon film.

in conclusion

(1) A sandwich-structured silicon-based anode was designed, using a flexible acetylene black coating instead of copper foil as the current collector. The active material can be tightly bonded between the acetylene black coating and the polyethylene film, buffering the volume change of silicon during the charge and discharge process, thereby improving the cycle performance.

(2) Adding VC to conventional LiPF6 electrolyte can form a stable SEI film on the silicon surface, thereby greatly improving the cycle performance of silicon film. Adding SO2Cl2 and LiBOB can improve the cycle stability of silicon film to a certain extent, while adding ET has no obvious effect.