Cite this Article
Liu Bonan, Lu Hao, Chu Geng, Luo Fei, Zheng Jieyun, Chen Shimou, Li Hong. Size effect of Si particles on the electrochemical performances of Si/C composite anodes. Chinese Physics B, 2018, 27(8): 088201  
Permissions
Size effect of Si particles on the electrochemical performances of Si/C composite anodes
Liu Bonan1, 2, Lu Hao1, 2, Chu Geng3, Luo Fei4, Zheng Jieyun1, Chen Shimou3, Li Hong1, 2, †
Beijing National Laboratory for Condensed Matter Physics, Institute of Physics, Chinese Academy of Sciences, Beijing 100190, China
School of Physical Sciences, University of Chinese Academy of Sciences, Beijing 100049, China
Key Laboratory of Green Process Engineering, State Key Laboratory of Multiphase Complex Systems, Institute of Process Engineering, Chinese Academy of Sciences, Beijing 100190, China
Tianmu Energy Anode Material Ltd. Co., Changzhou 213300, China

 

† Corresponding author. E-mail: hli@iphy.ac.cn

Project supported from the “Strategic Priority Research Program” of the Chinese Academy of Sciences (Grant No. XDA09010102).

Abstract

A series of Si/C composites were fabricated based on pitch and Si powders with particle sizes of 30, 100, 500, and 3000 nm. The size effects of the Si particles in the Si/C composites were investigated for lithium-ion battery anodes. The nanoscale Si and Si/C composites exhibited good capacity retentions. Scanning electron microscopy showed that exterior and interior cracks emerging owing to volume expansion as well as parasitic reactions with the electrolyte could well explain the performance failure.

PACS: 82.47.Aa;65.40.gk;82.45.Fk;62.23.Pq
Keyword:silicon-carbon composite;solid-electrolyte interphase;size effect;anode;volume expansion
1. Introduction

Since their first commercialization in the 1990s, lithium-ion batteries (LIBs) have gradually occupied the portable-electronics market owing to their advantages of a high energy density, good rate capability, long cycle life, and low self-discharge rate. Nowadays, lithium-ion batteries are widely regarded as promising for various applications including electric vehicles, large-scale energy-storage equipment, distributed mobile power supply, and other fields.[1–3]

As an indispensable component in LIBs, the anode plays a significant role in determining the key parameters, such as cycle stability, energy density, and safety performance. Extensive efforts have been devoted to the investigation of anode materials in the past few decades, and significant progress has been achieved. Coke is the first commercialized LIB anode material, which provides a reversible capacity of 200 mAh/g–250 mAh/g. Afterwards, graphite became an important anode material owing to the large capacity of 372 mAh/g.[4] The successful commercialization of graphite can be attributed mainly to the significant improvements in electrolyte systems.[5] Graphite materials correspond to over 95% of the LIB market owing to the superior stability and high reversibility. After decades of development, the practical capacity of graphite can reach 340 mAh/g–370 mAh/g, approaching its theoretical energy-density limit. Therefore, the development of novel anode materials is necessary for future battery designs.

Silicon-based materials are the most-promising anode materials owing to the extremely high theoretical capacity of 3580 mAh/g for the formation of Li15Si4.[6,7] Si-based anodes exhibit an appropriate operation voltage of 0.4 V–0.5 V versus Li+/Li, benign environmental properties, and low cost, which further increase their potentials. However, two main challenges have hindered wide applications of Si-based anodes. One of them is the volume inflation;[8,9] the volume expansion of silicon during lithiation and de-lithiation can reach up to 320%,[10] which is linearly proportional to the lithiation amount, irrespective of the particle size, morphology, or crystallinity.[11,12] The large stress and strain induced from volume changes would lead to cracking of active particles, exposure of fresh surfaces, and exfoliation of the electrode. The other challenge is the formation of an inhomogeneous and unstable solid–electrolyte–interphase (SEI) layer on silicon anodes during electrochemical cycling.[13] The SEI would continuously grow whenever a fresh surface is exposed owing to volume swell. The SEI on a silicon anode could grow to a thickness of several micrometers after dozens of cycles.[14] The unstable SEI hindered practical applications. In full cells, the continuous growth of unstable SEI rapidly consumes the limited lithium and electrolyte, leading to a fast capacity decay.[15,16]

In order to overcome these issues, several approaches have been proposed. The nano-silicon/carbon composite material, composed of well-dispersed nanoscale silicon in a carbon matrix, is one of the most-promising candidates.[17,18] The cracking mechanism and size effect of Si have been thoroughly studied in pure-Si systems. Surface cracking emerges during lithiation of silicon, which can be attributed to the large hoop tension in the outer layers of the particles.[19–23] Early studies confirmed that the critical size, below which the cracking did not occur, was ∼ 150 nm for crystalline Si and ∼ 870 nm for amorphous Si based on a transmission electron microscopy (TEM) characterization and mechanical simulations.[24,25] However, no extensive studies of the size effect of Si on the formation of cracks in Si/C composite materials have been reported. In this study, we synthesized a series of Si/C composites based on pitch and Si powders with sizes of 30, 100, 500, and 3000 nm. The electrochemical performances of these Si/C composite materials are studied. Scanning electron microscopy (SEM) measurements are employed to investigate the external and internal structures, and crack evolutions during different cycles.

2. Experimental methods
2.1. Synthesis

Si particles with sizes of 30 nm and 100 nm (Alfa Aesar) were laser-synthesized from vapor phase. Si particles with larger sizes of 500 nm and 3 μm were obtained through jet milling from 325-mesh-silicon powder (Alfa Aesar).

The precursors for the Si/C composites are silicon powders with different sizes of 30, 100, 500, and 3000 nm, and coal pitch. The pitch has a softening point of 200 °C and coking value larger than 70%. A mixture of Si and pitch with a weight ratio of 1:9 was sintered in a rotary kiln at 900 °C for 2 h to ensure uniform coating and carbonization of the pitch.

2.2. Electrochemical measurements

The pure-Si electrode contained 70 wt% of silicon, 20 wt% of CB, and 10 wt% of CMC. The Si/C composite electrodes were prepared with 93 wt% of Si/C active material, 2 wt% of carbon nanotubes as conductive additives, 2 wt% of Na–CMC, and 3 wt% of styrene-butadiene rubber (SBR) as binders. A slurry of the above mixture with deionized water was cast on a Cu foil current collector and then dried at 100 °C. Electrodes with dimensions of 8 mm × 8 mm were punched and dried under a vacuum at 110 °C overnight to remove traces of water and then introduced in an argon-filled glove-box. The weights of active material per electrode were approximately 0.3 mg/cm2 and 3 mg/cm2 for pure Si and Si/C, respectively. The electrolyte was a solution of 1-M LiPF6 in ethylene carbonate (EC) and dimethyl carbonate (DMC) (1:1 in volume) with 5% fluoroethylene carbonate (FEC) as an additive. The addition of FEC could help stabilize the SEI. A lithium foil was used as the counter electrode, while glass fiber was used as the separator. 2032-type coin cells were assembled in an argon-filled glove box.

Discharge and charge tests were performed on a Land BT2001 battery test system (Wuhan, China) in a voltage range of 0.005 V–1 V at various C-rates at room temperature (1 C corresponds to a current density of 0.6 A/g). Electrochemical impedance spectroscopy (EIS) measurements were performed at 1.0 V using an electrochemical workstation (CHI660B) with an amplitude of the AC signal of 5 mV in a frequency range of 4 MHz to 0.1 Hz.

2.3. Characterizations

The phase compositions of Si and Si/C were characterized using an x-ray diffractometer (D8 Advance, Bruker) equipped with a Cu–Kα radiation source in a scan range (2θ) of 10°–80° with an increment of 0.02°.

The size distributions of the 30-nm and 100-nm silicon particles were evaluated by TEM (FEI20). The samples were dispersed in alcohol and deposited on a carbon film.

After different numbers of cycles (1st, 10th, 20th, 50th, and 100th cycle) at the cut-off voltage of 1.0 V, the coin cells were disassembled in the glove box with a disassemble machine (MSK-110, MTI, China). The electrodes were washed at least three times with DMC and then dried in a vacuum box for at least 4 h. The morphologies of the surface and focused ion beam (FIB) prepared cross-section were investigated with an SEM (Hitachi S-4800) equipped with an energy-dispersive x-ray spectroscopy (EDS) setup. In order to avoid air contamination, all of the measurements and transfer processes were performed under an Ar atmosphere or vacuum environment.

3. Results and discussion
3.1. Morphology and electrochemical properties of the pure Si anode

Figure 1 shows the morphologies of the Si particles with different sizes. The Si particles with average sizes of 30 nm (Figs. 1(a) and 1(c)) and 100 nm (Figs. 1(b) and 1(d)) have narrow particle-size distributions and spherical shapes. The Si particles with larger sizes of 500 nm (Fig. 1(e)) and 3 μm (Fig. 1(f)) exhibit irregular shapes. The x-ray diffraction (XRD) patterns in Fig. 2 verify the phase purity of the four samples. The peak of the 30-nm Si is wider owing to the smaller grain and particle sizes.

Fig. 1. (color online) TEM images of Si particles with sizes of (a) 30 nm and (b) 100 nm. SEM images of Si particles with sizes of (a) 30 nm, (b) 100 nm, (c) 500 nm, and (d) 3 μm.
Fig. 2. (color online) XRD patterns of the Si particles with average sizes of 30 nm, 100 nm, 500 nm, and 3 μm.

The charging and discharging curves and electrochemical performances of the pure-Si samples with different particle sizes are shown in Fig. 3. The charging capacity attributed to the SEI was larger for Si with the sizes of 30 nm and 100 nm, owing to the larger surface area. In addition, it is worth noting that the de-lithiation platform in the discharging curve is oblique for the smaller Si particles owing to the size effect on the surface energy.[26] Their initial discharge capacities are approximately equal, with values of approximately 3200 mAh/g–3300 mAh/g. After 100 cycles, the discharge capacities of the Si samples with the particle sizes of 30 nm, 100 nm, 500 nm, and 3 μm decrease to 2454, 2189, 1483, and 279 mAh/g, respectively. The Si particles with smaller sizes exhibit significantly better capacity-retention behaviors. This result is consistent with a previous report by Li et al.[27] They reported that the volume variation of single nanometer alloy particles was significantly smaller than that of single large-size alloy particles during discharge–charge cycles. Moreover, the higher plasticity and deformability of nanomaterials prevent electrode cracking and pulverization.

Fig. 3. (color online) (a) Charging and discharging curves of the first cycle and (b) cycling performances of the Si particles with different sizes at a current density of 0.2 A/g in the voltage window of 0.005 V–1 V.
3.2. Morphology and electrochemical properties of the Si/C composites

The XRD patterns of the Si/C composites are typical patterns of a simple physical mixture of components: silicon and pitch coke (Fig. 4); no silicon–carbide formation was detected upon heat-treatment.

Fig. 4. (color online) XRD patterns of the Si/C composites with Si particle sizes of 30 nm, 100 nm, 500 nm, and 3 μm.

Figure 5 reveals the morphologies of the Si/C composites containing Si particles with different sizes. It is worth noting that after the calcination process with pitch melting, the Si/C composites have microscale sizes in the range of 10 μm–15 μm with “potato-like” shapes; they have a wide size distribution. The pristine Si particles are difficult to be identified, which implies that the pitch-derived carbon successfully wraps them; the Si particles are well dispersed in the carbon matrix.

Fig. 5. (color online) SEM images of the Si/C composites with Si particle sizes of (a) 30 nm, (b) 100 nm, (c) 500 nm, and (d) 3 μm.

The electrochemical performances of the Si/C composites were evaluated; the results are shown in Fig. 6. It should be mentioned that the areal capacity of the electrode is 2.5 mAh/cm2, close to the practical application level. The charging and discharging curves demonstrate a combination of Si and soft carbon, particularly in the discharging curve.[28] In the dQ/dV curves (Fig. 6(b)), the peaks at ∼ 0.42 V are higher for the 500-nm and 3-μm Si/C, consistent with the conclusion from Fig. 3(a), indicating that the size and structure of silicon have not been altered during the sintering and milling of Si/C. The capacity variations of these Si/C composites as a function of the cycling time are shown in Fig. 6(c). All of the Si/C composites deliver initial discharge capacities of approximately 600 mAh/g, corresponding to 12 wt% of silicon, considering that the theoretical capacity of Si is 3200 mAh/g and that the theoretical capacity of the pitch-derived carbon is 250 mAh/g. The cycling performance deteriorates with the increase of the Si particle size. The initial discharge capacities are 605, 586, 592, and 584 mAh/g for the 30-nm, 100-nm, 500-nm, and 3-μm Si/C, with corresponding capacity retentions of 87.5%, 81.4%, 67.2%, and 48.1%, respectively, after 100 cycles. In line with the pure Si anode, the Si particles with smaller sizes in the composites deliver better electrochemical performances. A better capacity-retention capability is accompanied with a high Coulombic efficiency, as demonstrated in Fig. 6(d). The average Coulombic efficiencies of the 10th–100th cycles for the 30-nm, 100-nm, 500-nm, and 3-μm Si/C are 99.3%, 98.9%, 98.6%, and 98%, respectively. In summary, the Si/C composite with a smaller Si particle size has a better capacity retention upon the discharge–charge processes.

Fig. 6. (color online) (a) Charging and discharging curves of the first cycle and (b) dQ/dV curves corresponding to the de-lithiation process. (c) Cycling performances and (d) Coulombic efficiencies of the Si/C composites as a function of the number of cycles at a current density of 0.12 A/g in the voltage window of 0.005 V–1 V.
3.3. Morphology evolutions of the Si/C composites

It is known that Si suffers from a large volume expansion during lithium insertion, which may lose contact with the conductive materials and lead to a high polarization. The exposed fresh surface-active materials can promote side reactions with the electrolytes, consuming the electrolyte and causing a low Coulombic efficiency. In this study, we track the surface morphology evolutions of the electrodes after different numbers of cycles (1st, 20th, 50th, and 100th cycle), as shown in Fig. 7. The 30-nm and 100-nm Si/C composites remain flawless after the first cycle. With the increase of the number of cycles, the number of cracks generated in the electrodes increases. For the 500-nm and 3-μm Si/C composites, cracks can be observed in the electrodes just after the first cycle. After 20 cycles, the particles are completely torn by the large volume expansion/contraction during the cycling. It is worth noting that the electrodes of the 500-nm and 3-μm Si/C composites were pulverized when we disassembled the cells.

Fig. 7. (color online) Surface morphology evolutions of the Si/C composites after different numbers of cycles at the cut-off voltage of 1.0 V: (a) 30-nm, (b) 100-nm, (c) 500-nm, and (d) 3-μm Si/C. The scale bar corresponds to 5 μm.

The Si particles are wrapped by the pitch-derived carbon; it is necessary to understand the internal cracking within particles. Using FIB techniques, we obtained the cross-sections of the Si/C composites (Fig. 8). For the 30-nm Si/C composite (Fig. 8(a)), the silicon particles slightly aggregate inside the carbon matrix owing to their high specific surface area. This composite can be divided into an Si-rich part and C-rich part. Based on the SEM image, there are no obvious cracks after the 1st cycle. After 20 cycles, it has cracks, consistent with the surface morphology. It is worth mentioning that the cracks emerge in the C-rich part rather than in the Si-rich part. Even after prolonged cycles, large cracks were not observed in the Si-rich part. The Si-rich part was blurred after 100 cycles, owing to a large amount of SEI, as discussed below. The 100-nm Si/C composite has a similar behavior to that of the 30-nm Si/C composite (Fig. 8(b)); however, a larger cleavage and smaller fractured-particle size are observed. The cracks did not propagate along the Si/C interface. For the 500-nm Si/C composite, no Si aggregation is observed owing to the large size of the silicon particles. Cracks appear after the first cycle, but do not exacerbate at the Si/C interface; no obvious cracks in silicon were observed until the 20th cycle. The width of the cracks could exceed 500 nm after 50 cycles (red circle). Inside the cleavage, SEI grows inside the crack after prolonged cycles. The surrounding carbon matrix may decrease the hoop tension, as reported by Huang et al. who revealed that the carbon matrix can alleviate the volume expansion of SnO.[29] The 500-nm Si/C composite is rapidly torn; silicon particles were fractured and detached from the matrix (Fig. 8(d)). This well explains its lower capacity retention and lower Coulombic efficiency, as demonstrated in Fig. 6. Larger silicon particles yield a larger stress on the carbon matrix leading to a faster fracture. It is worth noting that even when the silicon size is as small as 30 nm, the composite still fails under the large stress.

Fig. 8. (color online) Internal morphology evolutions of the Si/C composites after different numbers of cycles at the cut-off voltage of 1.0 V: (a) 30-nm, (b) 100-nm, (c) 500-nm, and (d) 3-μm Si/C. The scale bar corresponds to 5 μm.
3.4. SEI formation and impedance

Energy-dispersive spectrometry (EDS) is employed to analyze the C, Si, and F elemental distributions for the 30-nm Si/C composite after 20 and 100 cycles (Fig. 9). It reveals that the points A and C are C-rich areas, while the points B and D are Si-rich areas. After 20 cycles, there is a small F-element signal at the point A, suggesting almost no formation of SEI. As the electrolyte contains 5% of FEC and the electrode was rinsed thoroughly with DMC, the F element should mainly originate from the SEI. After 100 cycles, the whole F content increases from 1.3% to 9.95%, indicating the formation of a large amount of SEI. In addition, for the Si-rich part inside the particle, the content of F significantly increases to 11.08%. As shown in Fig. 8, the Si-rich part blurred after 100 cycles. Once the composite material cracks, the electrolyte would permeate inside the particles leading to a continuous growth of SEI. The EDS results are shown in Table 1.

Fig. 9. (color online) Cross-section SEM images of the 30-nm Si/C composite after (a) 20 cycles and (b) 100 cycles. The scale bar corresponds to 5 μm.
Table 1.

EDS results based on the area of Fig. 7.

.

The EIS with Nyquist plots of the Si/C composites after different numbers of cycles at the cut-off voltage of 1.0 V, and corresponding Rct and RSEI evolutions obtained using an equivalent-circuit fitting are presented in Fig. 10. Overall, after the activation process with a decreasing Rct in the first 20 cycles, the Rct gradually increases; the 3-μm Si/C composite has the largest Rct of 390 Ω; that of the 30-nm Si/C composite is 150 Ω at the 100th cycle. This suggests that a worse electrical connection occurs, particularly at the large particle size owing to the crack formation. For the SEI, the RSEI increases with the number of cycles, indicating a continuous formation of SEI in these composites. The 30-nm Si/C composite has the smallest RSEI, which can well explain its high capacity and good cycling-duration performance.

Fig. 10. (color online) EIS of the Si/C composites after different numbers of cycles at the cut-off voltage of 1.0 V: (a) 30-nm, (b) 100-nm, (c) 500-nm, and (d) 3000-nm Si/C. (e) Charge-transfer resistance evolutions as a function of the number of cycles for the composites. (f) SEI resistance evolution as a function of the number of cycles for the composites.
4. Conclusions

Si/C composites were fabricated by sintering petroleum pitch and Si with particle sizes of 30, 100, 500, and 3000 nm. Both pure Si and Si/C composites with the smallest particle size (30 nm) had the best capacity retentions as well as the smallest RSEI increments. Ex-situ SEM assisted with the FIB technique demonstrated that the pitch-derived carbon could relieve the strain and stress of Si during cycling. Cracks were easily generated in the large particles during cycling; the cracks did not propagate along the Si/C interface, owing to the interplay between the interface strength and hoop tension. Once the composite cracks, electrolyte would permeate inside active particles and lead to a continuous growth of SEI, which is detrimental for the electrochemical performances of the Si/C composite materials.

Reference
[1] Bridel J S Azaïs T Morcrette M Tarascon J M Larcher D 2010 Chem. Mater. 22 1229
[2] Kierzek K Machnikowski J Béguin F 2014 J. Appl. Electrochem. 45 1
[3] Luo F Liu B Zheng J Y Chu G Zhong K F Li H Huang X J Chen L Q 2015 J. Electrochem. Soc. 162 A2509
[4] Flandrois S Simon B 1999 Carbon 37 165
[5] Xu K 2004 Chem. Rev. 104 4303
[6] Hatchard T D Dahn J R 2004 J. Electrochem. Soc. 151 A838
[7] Li J Dahn J R 2007 J. Electrochem. Soc. 154 A156
[8] Pereira-Nabais C Światowska J Chagnes A Ozanam F Gohier A Tran-Van P Cojocaru C S Cassir M Marcus P 2013 Appl. Surf. Sci. 266 5
[9] Philippe B Dedryv‘ere R Allouche J Lindgren F Gorgoi M Rensmo H Gonbeau D Edström K 2012 Chem. Mater. 24 1107
[10] He Y Yu X Q Li G Wang R Li H Wang Y L Gao H J Huang X J 2012 J. Power Sources 16 131
[11] McDowell M T Lee S W Nix W D Cui Y 2013 Adv. Mater. 25 4966
[12] Obrovac M N Chevrier V L 2014 Chem. Rev. 114 11444
[13] Zheng J Y Zheng H Wang R Ben L B Lu W Chen L W Chen L Q Li H 2014 Phys. Chem. Chem. Phys. 16 13229
[14] Luo F Chu G Xia X X Liu B N Zheng J Y Li J J Li H Gu C Z Chen L Q 2015 Nanoscale 7 7651
[15] Dupre N Moreau P De Vito E Quazuguel L Boniface M Bordes A Rudisch C Bayle-Guillemaud P Guyomard D 2016 Chem. Mater. 28 2557
[16] Nguyen D T Kang J Nam K M Paik Y Song S W 2016 J. Power Sources 303 150
[17] Holzapfel M Buqa H Krumeich F Novák P Petrat F M Veit C 2005 Electrochem. Solid State Lett. 8 A516
[18] Terranova M L Orlanducci S Tamburri E Guglielmotti V Rossi M 2014 J. Power Sources 246 167
[19] Zhao K Pharr M Wan Q Wang W L Kaxiras E Vlassak J J Suo Z G 2012 J. Electrochem. Soc. 159 A238
[20] Ma Z Li T Huang Y L Liu J Zhou Y Xue D 2013 RSC Adv. 3 7398
[21] Zhao K Pharr M Cai S Vlassak J J Suo Z G 2011 J. Am. Ceram. Soc. 94 226
[22] Huang S Fan F Li J Zhang S Zhu T 2013 Acta Mater. 61 4354
[23] Kalnaus S Rhodes K Daniel C 2011 J. Power Sources 196 8116
[24] Liu X H Zhong L Huang S Mao S X Zhu T Huang J Y 2012 ACS Nano 6 1522
[25] McDowell M T Lee S W Harris J T Korgel B A Wang C Nix W D Cui Y 2013 Nano Lett. 13 758
[26] Li H Wang Z X Huang X J Chen L Q 2008 Physics 37 416 in Chinese http://www.wuli.ac.cn//CN/abstract/abstract31231.shtml
[27] Li H Huang X J Chen L Q Wu Z G Liang Y 1999 Electrochem. Solid State Lett. 2 547
[28] Lu H Liu B N Chu G Zheng J Y Luo F Qiu X P Li H Liu F Feng S N Chen W Li H Chen L Q 2016 Energy Storage Science and Technology 5 109
[29] Zhang L Q Liu X H Liu Y Huang S Zhu T Gui L Mao S X Ye Z Z Wang C M Sullivan J P Huang J Y 2011 ACS Nano 5 4800