Abstract

Si is essential as an active material in Li-ion batteries because it provides both high charge and optimal cycling characteristics. A composite of Si particles, Cu particles, and pure H2O was realized to serve as an anode active material and optimize the charge–discharge characteristics of Li-ion batteries. The composite was produced by grinding using a planetary ball mill machine, which allowed for homogenous dispersion of nanoscale Cu3Si as Si–Cu alloy grains and nanoscale Si grains in each poly-Si particle produced. Furthermore, some Si particles were oxidized by H2O, and oxidized Si was distributed throughout the composite, mainly as silicon monoxide. As a result, each Si particle included silicon monoxide and conductive Cu3Si materials, allowing for effective optimization of the recharging and charge-discharge characteristics. Thus, a new and simple process was realized for synthesizing a Si active material composited with silicon oxides, including silicon monoxide. This Si-rich conductive material is suitable as an anode for Li-ion batteries with high charge and optimized cycling properties.

1. Introduction

Owing to its high theoretical Li storage potential of about 4000 mAh g−1 [1], Si is one of the most attractive anode materials for Li-ion batteries (LIBs). However, Si undergoes frequent drastic changes in volume during charge-discharge cycling, causing degradation of Si materials and drastically decreasing ionic and electronic conductivities, which has prevented Si anodes from being utilized in Li-ion secondary batteries [2]. To overcome this problem, various modifications, such as the use of nanostructured Si anodes, the synthesis of composites with other materials, and the use of carbon coatings, have been suggested [35]. The use of Si monoxide as an alternative anode material has been proposed [611], as Si monoxide exhibits little change in volume and high conductivity. However, the capacity of Si monoxide is only about 1200 mAh g−1, which is lower than that of other potential materials [12, 13]. Numerous attempts have been made to find ways to compensate for this by combining Si monoxide with other materials [1417].

This study considers a technique that maintains the high ionic and electronic conductivities of Si by preventing cracking during volume changes, with as little effect as possible on the capacity of the material. Namely, particles of an active material based on Si were formed into the core of a suitable conductive material and the Si crystal structure was controlled by compositing Si particles with other materials. To achieve this, we designed a composition including Si, Si oxide, and a conductive material, which could be employed as an anode electrode in LIBs. In this study, to synthesize a composite of Si and other materials, we used a grinding process to combine Si with H2O as an oxidant and Cu as a conductive material. The grinding method was considered the most likely method to achieve a composite containing a homogeneous distribution of nanoscale Si and conductive materials. It was expected that a Si oxide with a low Si oxidation number would be formed by grinding with H2O, resulting in the transfer of oxygen atoms the Si and Si particles. Moreover, a composite material containing Si oxide, Cu, and Si is expected to yield good performance with regard to charge-discharge characteristics.

2. Experimental

2.1. Synthesis of Composites of Si, H2O, and Cu

In this study, we developed a grinding method in which Si particles (average diameter of 4 μm, purity of 99.999%; Kojundo Chemistry Laboratory Co., Ltd., Japan) and Cu particles (average diameter of 4 μm; Kojundo Chemistry Laboratory Co., Ltd., Japan) as a conductive material were ground with pure H2O in the mill pots of a planetary ball mill (Fritsch Pulverisette-7) to produce a composite material. A mixture of approximately 2 g was put in a zirconia mill pot with a 45 cm3 inner volume with 7 zirconia balls of 15 mm diameter and subjected to grinding using the planetary ball mill, as shown in Figure 1, at approximately 600 rpm in air. The mill pots rotated individually on a rotating disk. The crushing conditions were varied by changing the amounts of Cu, pure H2O, and Si particles, as well as the grinding time. Table 1 shows the amounts of Si, H2O, and Cu used in this study to prepare composites by grinding. The “” values in Table 1 indicate the ratio of oxide atoms in the prepared composite relative to Si. The amount of Cu was constant based on the conditions for grinding Si and CuO [18]. Crushing and reconstruction of materials in the mill pot was expected to result in the formation of a composite containing Si, Cu, and oxidized species. Oxidation reactions that result in the transfer of an oxygen atom from H2O to Si would produce a Si suboxide in the composite with Si and Cu. Further, residual Cu can serve as an electrically conductive material. Such composites synthesized from Si, H2O, and Cu could then be employed as the active material in LIB anodes.

2.2. Characterization of Composite Material

Scanning electron microscopy (SEM; Hitachi High-Technologies Corporation, Japan) was used to examine the morphology of the composite particles. A composite-coated electrode was cut using a Ga focused ion beam (FIB; Hitachi High-Technologies Corporation, Japan) at an accelerating voltage of 5 kV. Scanning transmission electron microscopy (STEM; Hitachi High-Technologies Corporation, Japan) was carried out at an accelerating voltage of 200 kV to image cross sections of the electrode. The atomic distribution of Si, oxides, and Cu was measured using energy dispersive X-ray spectroscopy (EDX; Hitachi High-Technologies Corporation, Japan). The crystallization state of the ground composite was examined using X-ray diffraction (XRD; Rigaku Co. Japan). Electron energy-loss spectroscopy (EELS; Hitachi High-Technologies Corporation, Japan) at 200 kV was used to measure the distribution of oxides in the composite at a resolution of 0.5 eV. The EELS map was constructed from spectral images in the energy-loss region of 200–2200 eV, as detected using multiple charge-coupled devices. High-resolution Si2p X-ray photoelectron spectroscopy (XPS; Bruker) with a monochromatic AlKα X-ray source, an analysis range of 0.62 mm in diameter, and a detection angle of 45° was used to determine the oxidation number of Si in the composite particles. The composite particles were ground uniformly on a XPS sample holder, and measurements were obtained for the flattest section.

2.3. Anode Preparation and Electrochemical Testing

Anode electrodes for LIBs were prepared using the composites of Si, H2O, and Cu as follows. Composite particles were mixed with a binder composed of polyamic acid (Ube-Kousan KK, Japan) and acetylene black (AB; Denkikagaku Kogyo KK, Japan) as a conductive material in a 1-methyl-2-pyrrolidone (NMP) solution. The weight ratio of composite particles : binder : AB was 70 : 20 : 10. To form electrodes, this slurry was cast onto a Cu foil and dried at 70°C for 30 min in air. The coated electrodes were cut to a size of 10 mm with a 78.5 mm2 area. The electrode thickness was within the range of 40–50 μm. The electrodes were further sintered at 650°C under vacuum for 3 h and then pressed at 200 kgf cm−2. The specific capacity was calculated according to the weight of Si in the composite particles.

Electrochemical testing of the composite electrodes was conducted using two-electrode test coin cells (2032-type, Housen, Japan) with caulked metal cups and a gasket to hold the electrode assembly and a separator. The coin cells were assembled in an Ar-filled glove box using a 1 M lithium hexafluorophosphate (LiPF6) electrolyte in a mixed solution of ethylene carbonate (EC), diethylene carbonate (DEC), and dimethyl carbonate (DMC) (60 : 25 : 15, v/v). An Al foil coated with a LiCoO2 film (320 μm thickness) with a capacity of 4.5 mAh cm−2 was employed as the cathode counter electrode, with a polyethylene/polypropylene/polyethylene multistacked film separator (40 μm thickness). The cathode electrode, separator, and anode electrode were not pressed before assembly in the test coin cell and were held in place by a leaf spring. The electrochemical performance of these two-electrode test coin cells was evaluated using a constant current charge-discharge cycling test in the voltage range of 1.6–4.2 V, with a current density of 0.1 mA cm−2 at room temperature.

3. Results and Discussion

3.1. Characterization of the Composite

In this study, we obtained composite particles by grinding Si, H2O, and Cu (1.88 g Si, 0.12 g H2O, and 0.28 g Cu; Table 1) in a planetary ball mill machine for a total of 3 h to obtain a homogeneous distribution of each component (Si, O, and Cu) in the composite. The molar ratio of Si : O in the composite was controlled at 10 : 1. The SEM images in Figure 2 reveal the composite particles obtained after grinding Si, H2O, and Cu. The composite particles, which are similar in size to the bare Si particle, appear to have small particles attached along their circumferences. The composite particles were from 0.1 to 3.8 μm in size with an existence probability of 95%, and the medium diameter was 1.6 μm. Further, the average tapping density measured for the prepared sample was 2.19 g cm−3, which is comparable to the tapping density of Si (approximately 2.3 g cm−3).

An electrode coated with a mixture of composite particles, conductive materials, and a binder was cut using a Ga FIB to analyse electrode cross sections by STEM. The dark field image of the sliced anode electrode (Figure 3) shows dark grey areas corresponding to Si components and white areas corresponding to Cu or Cu components. Thus, it was revealed that the Si composites were homogeneously mixed with Cu materials. Punctate white regions were also observed in the composite materials, and we surmised that the composites were synthesized from a mixture of materials based on Si, H2O, and Cu. EDX was used to measure the atomic distribution of Si, oxides, and Cu (Figure 4) in the composite. These results show that the material obtained following the grinding process contains both Si and Cu nanograins.

Figure 5 shows a high-resolution STEM image at 2000 K magnification of the area of the composite indicated by a white circle in the inset. The sample composition was observed as aggregates of nanoscale grains based on individual Si and Si–Cu composite materials. The bright grey areas, indicated by white arrows in Figure 5, are occupied by Si materials, whereas the dark grey areas, indicated by black arrows, contain Cu materials. The directions of these arrows indicate the orientation of each Si or Cu crystal lattice. Thus, high-resolution STEM imaging confirms that each grain has a random crystal orientation and that the composite mainly comprises poly-Si and Si–Cu composite nanograins.

The crystallization state of the ground composite was examined using XRD, as shown in Figure 6. Only Si and Cu particles were observed before the grinding process, whereas in the composite obtained after grinding, crystal patterns were observed that correspond to Si and Cu3Si as a Si–Cu alloy [19]. Most oxygen atoms from H2O were transferred to Si by the grinding process; on the other hand, Cu did not react with H2O directly. These results confirm that the material obtained by grinding is a composite constructed of aggregated nanoscale grains of Si, Si oxidation materials, and Cu3Si materials.

The distribution of oxides in the composite was measured using EELS. The EELS map obtained from STEM images is shown in Figure 7. Owing to oxidation reactions initiated by the grinding process, most oxygen atoms from H2O were incorporated with Si into Si oxides. The EELS map in Figure 7, which is constructed from the spectral data at energy losses between 532 and 570 eV, implies that the distribution of oxide species, indicated by white dots, is not homogeneous in the composite. Thus, oxide species are concentrated near the surface of the composite particle, although some oxides also exist inside the particle.

The Si oxidation number of the composite particles was determined using high-resolution Si2p XPS spectroscopy. The binding energies of the peaks in the XPS spectra were extrapolated by fitting with the Voigt profile [2022]. The peak corresponding to carbon contaminants was employed as a reference for the binding energy. The XPS results shown in Figure 8 indicate that the composite contains a variety of Si ions with different oxidation numbers. The Si2p peaks located at higher binding energies (~100–104 eV) mainly correspond to Si-suboxide species: Si4+, Si3+, Si2+, and Si1+. The content ratios of the different Si ions in the composite particles were calculated by deconvolution of the peak, which was calibrated against the C1s signal. In addition, each Si oxidation number was evaluated by calculating the chemical shift distribution based on the results shown in Figure 8. Si in the composite particles is mainly Si0 (Si); however, the distribution of the binding energies indicates that the composite also contains some oxidized Si species: Si4+, Si3+, Si2+, and Si1+. This finding confirms the existence of a semisilicate, similar to Si monoxide, near the surface of the particles. Approximately 75% of the XPS spectrum corresponds to Si and Si1+, suggesting that the composite has a strong metallic character. Nearly 80% of all other oxidized species have oxidation numbers of Si2+ and Si3+. The formation of Si monoxide or semimonoxide is expected owing to the controlled amount of H2O added before the grinding process. These findings are evidence of the synthesis of Si monoxide or submonoxide, which are potentially useful as active materials in LIBs, based on oxidization reactions during grinding with H2O.

3.2. Electrochemical Performance of the Composite

The electrochemical performance of the composite anode was evaluated in a 2032-type test coin cell using a constant current charge-discharge cycling test in the voltage range of 1.60–4.16 V with a current density of 0.1 mA cm−2 at room temperature. The state of charge was controlled at 100% in this study. As shown in Figure 9, the 1st charge-discharge capacity of the composite is more favourable than that using Si particles as a reference anode. The capacity was calculated using the weight of Si included in the anode of the coin cell. The 1st charge and discharge capacities for the composite were 3384 and 3160 mAh g−1, respectively, whereas those for Si particles were 3871 and 3198 mAh g−1, respectively.

The recharging capacity results indicate that the charge and reversible discharge capacities depend on the composition of the active material. The composite obtained by grinding Si, H2O, and Cu demonstrated a 1st coulombic efficiency of 94.1% (Figure 10, red), whereas a 1st coulombic efficiency of 82.9% was observed when the anode only contained Si particles (Figure 10, blue). When the active material only contains Si, the 1st coulombic efficiency approaches 83% owing to formation of an irreversible solid electrolyte interphase (SEI), a chemical compound consisting of Li, Si, and solvent of electrolyte, on each Si particle, which does not work as an active material [1, 3]. Furthermore, Si crystals form amorphous or quasiamorphous Si phases after discharging Li ions, which have a lower capacity than the theoretical capacity of crystalline Si [5].

The composite produced by grinding Si, H2O, and Cu was successfully employed as an anode to occlude Li ions. The composite is constructed of aggregates of Si, Si oxidation products, and Cu3Si nanoscale particles, as indicated in Figure 5. We surmise that the Si nanoscale grains in the composite help to occlude Li ions with high coulombic efficiency during the first charge-discharge cycle; moreover, the Cu3Si nanoscale grains act as an electrically conductive material. Further, to obtain good coulombic efficiency, the mixture of Si and Si monoxide prevents irreversible formation of an SEI [2325]. In such mixtures, the optimized recharging efficiency and the conductivity of Si monoxide are lower than those of other Si oxides [26].

The cycling characteristics of the composite anode and a reference anode employing only Si particles are compared in Figure 11. The composite demonstrated a reversible capacity of over 3000 mAh g−1 after 100 cycles. Moreover, the capacity retention was over 99.9% for the 2nd cycle and 91.6% for the 100th cycle (Figure 11). In contrast, the bare Si particle anode had a low capacity retention of less than 50% over 40 cycles.

The monoxide or submonoxide Si species in the composite form near the surface, as shown in Figure 6. Thus, the structure of the composite was visualized as a Si active material covered by an oxide film. The optimization of the cycling characteristics by covering Si particles with an oxide film has been reported [26, 27]. The capacity retention was clearly found to be optimized by covering the active materials of the LIB anode with a thin film coating. It is believed that the film covering each Si particle plays a buffering role by preventing cracking following drastic changes in volume with numerous repetitions of the charge-discharge cycle. Based on our results, the composite has a structure consisting of aggregated Si, Cu nanoscale grains, and Si oxidation products, which accumulate as Si oxides at the surface of the composite and allow for optimization of the electrochemical characteristics of LIB cells, as shown in Figure 12.

The prepared composite sample exhibited nanostructures corresponding to both a Si matrix and Cu3Si grains, which may be partly responsible for the improvement in battery performance, as reported in other studies on the preparation of nanostructured samples by various methods. However, we consider that there are other possible reasons for the improved performance of our composite sample. Notably, the extent of oxidization of the Si grains played an important role in increasing the coulombic efficiency, as revealed when we varied the ratio of H2O to Si in the sample. Figure 13 shows the relationship between the initial coulombic efficiency and the molar ratio of Si to H2O used during composite preparation. Without added H2O, the coulombic efficiency was as low as 75%, but this value increased rapidly to 93.7% for a Si to H2O ratio of 10 : 1. Further increasing the amount of H2O led to a decrease in the coulombic efficiency. We postulated that higher amounts of H2O during the mechanochemical reaction between Si, H2O, and Cu would result in the production of a highly oxidized state of Si, with a higher ratio of SiO2 in the obtained composite. Highly oxidized Si species such as SiO2 tend to form stable Li silicates during the battery charging operation. When a Si to H2O ratio of 10 : 1 was used for sample preparation, as shown in Figures 38, the obtained reaction product was SiOx, where is around 1 and much smaller than 2, as XPS analysis showed the formation of an amorphous layer around the crystalline Si grains. During lithiation, SiOx with a low oxidation state will have a low tendency to react with Li to form silicates. On the other hand, the large increase in conductivity reported to be associated with low oxidation states of Si [26] will make the amorphous SiOx layer more conductive than the Si matrix grains. Moreover, this layer acts as a physical buffer to reduce volume changes, thereby increasing efficiency.

As an excellent conductor, Cu3Si has been widely studied for the improvement of Si anode performance through various modifications, such as coated layers on Si grains, and increased cycling efficiencies have been reported [2831]. As shown in Figure 3, Cu3Si nanoparticles with sizes of less than 10 nm are dispersed uniformly among the Si grains in our composite, which increases its conductivity. In addition to contributing to electronic conductivity, Cu3Si may also serve as a good catalyst for increasing the cycling efficiency. There have been several reports on the catalytic effects of Cu3Si in producing chlorosilanes from the reaction between CH3Cl and Si [32]. Harper et al. have reported room-temperature oxidation of Si catalysed by Cu3Si [33]. These studies indicate the high ability of Cu3Si to catalyse the oxidation of Si to form chlorides or oxides. Although there have been no direct studies reporting the oxidation of Li silicides in anode materials, we believe that, in our prepared sample, the Cu3Si nanoparticles dispersed uniformly around the Si grains, which exist as Li silicides after charging, will catalyse the oxidation of Li silicides and facilitate Li release, so that the rate of capacity fading is reduced. In addition to catalysis by Cu3Si, the synergistic effects of the nanostructure of Si grains and the amorphous oxidized layers formed around the Si grains are considered to contribute to the observed high performance, with a capacity of over 3000 mAh g−1 maintained after 100 cycles. The synergistic effects result from the mechanochemical approach, which is generally studied for stoichiometric reactions. However, in this case, nonstoichiometric ratios of H2O and Cu were used to successfully prepare Si anode materials, expanding our understanding of mechanochemical phenomena.

4. Conclusion

Though it is necessary to employ Si as an active material to achieve both high charge and good cycling characteristics for charge-discharge performance, it is difficult to control Si active materials. We created a composite of active materials as an anode electrode for LIBs by compounding Si particles with H2O and Cu particles using a grinding process. Oxidation-reduction reactions between Si and H2O were activated by this process, resulting in Si oxidation. Thus, a composite containing aggregates of Si, Si oxidation products, and Cu3Si nanoscale grains was achieved.

The Si oxidation number of the Si oxidation products could be controlled by the amount of H2O added during grinding. Si monoxide was located near the surface of the composite, as revealed by EELS and XPS measurements. Therefore, the composite acts as an active material with high capacity and excellent charge-discharge properties over many cycles. Thus, we consider that further drastic improvement may be obtained from synergy effects when two or more actions are simultaneously introduced with only one operation. And this synthesis method for new anode materials and the battery, which has this anode, a commercial cathode, and an electrolyte, will be available for practical use in the commercial battery market.

While optimizing our treatment conditions for preparation of the Si composite, we are considering different applications of nonstoichiometric reactions to other similarly intricate issues. For example, we expect to be able to synthesis a cathode composite having more than three Li atoms in a molecule via the grinding process.

Conflicts of Interest

The authors declare that they have no competing interests.

Acknowledgments

The authors kindly thank DOWA Holdings Corporation for their help in the construction and electrochemical measurements of the composite employed in this study. The authors would like to thank Editage (https://www.editage.jp) for English language editing.