Effect of surface modification on electrochemical performance of nano-sized Si as an anode material for Li-ion batteries

Abstract

Silicon is one of the most promising anode materials for lithium-ion batteries. To solve the problems associated with the great volume expansion of Si during lithium storage, nano-sized Si particles are generally employed. However, their high surface activity is likely to trigger considerable electrolyte decompositions at low potential, thus the surface of these Si nano-particles need further chemical modifications. In this paper, three kinds of functional groups were grafted onto the surface of Si particles by different chemical treatments. X-ray photoelectron spectroscopy (XPS) studies proved that the structure of solid electrolyte interface (SEI) film formed on the surface of Si nano-particles depends greatly on the surface modification strategy. Electrochemical characterizations like electrochemical impedance spectroscopy (EIS), cyclic voltammetry (CV) etc., also verified the distinct effects of these functional groups on the surface of nano-sized Si electrodes. The relationship between surface functional groups and electrochemical performance of the nano-Si anode material was addressed.

---

1. Introduction

For Li-ion batteries, silicon is a very promising anode material with a theoretical specific capacity (∼3580 mA h g⁻¹) much higher than that of the popular anode today (graphite, ∼372 mA h g⁻¹).^1,2 However, a crucial problem is the large volume changes (∼300% volume expansion) upon lithium alloying/dealloying, which lead to cracks within Si particles and loss of electrical contacts with copper foil.^3–5 Thus, the electrodes may exhibit a big electrochemical polarization and irreversible capacity fading along cycles.^6,7 In order to avoid these shortcomings, nano-structured Si was always adopted.^8,9 There are at least two advantages. First, nano-particle could shorten the transmission path of Li-ions to reduce potential polarization; second, it could provide adequate accommodation surface to efficiently release the stress from drastic volume changes of crystal lattice and then prevent degradation of electrodes structure.^10,11 At first alloying with lithium, the surface of silicon undergoes a lot of chemical side-reactions with electrolyte to form a deposition film, which is so called the solid electrolyte interphase (SEI) film.^12–14 The SEI film may be composed of lithium carbonate (Li₂CO₃), lithium alkyl carbonates (ROCO₂Li), ROLi, lithium fluoride (LiF) and other solid materials. The SEI can prevent further parasite chemical reactions on surface of Si-particles and ensure the reversibility of the electrodes.^15,16 At the same time, the formation of SEI may consume a large amount of lithium and other electrolyte components in a lithium-ion battery, and consequently it may lead to the high irreversible capacities. Moreover, the over-thick SEI also can block transmission of Li-ion on interface and decrease electrochemical activity.^17,18 Therefore a favorable surface of nano-particles plays an important role for the Si-based anodes.

On the other hand, the chemical activity of Si nanoparticles surface will inevitably give rise to silicon dioxide layers on the exposure in air. This oxide layer may be as thin as 1–3 nm.¹⁹ The thickness also varies in accordance with conditions. At first alloying, the formation of SEI is partly generated from the reaction of the oxide layer and electrolyte.^20,21 and the oxide layer is turned to a variety of lithium compounds, which take effects along the continuous lithium uptake/release cycles.²² An XPS study found that the oxide layer was partly transformed into Li₄SiO₄ and Li₂O during the first discharge process.²³ Furthermore, the thickness of oxide layer also has a great influence on the quality of SEI film.²⁰ HF-etching appeared an effective way to adjust the thickness of oxide layer, and thinner dioxide layer could increase the first reversible capacity. In addition, this kind of oxide layer tends to react slowly with Li-ion during cycling and cause capacity fading.²⁴ In contrast with the deleterious effects of the oxide layer as mentioned above, the work by Song et al. reported demonstrated that a fine oxide layer can improve cyclability by the deposition an organic silane (Si–O–Si) layer on surface of silicon, where the Si–O–Si network can passivated particles surface against the electrolyte.²⁵ Ryu et al. also exhibited similar effect of silanes on silicon anode when silanes were used as additives in electrolyte are effective to form stable SEI of Si–O–Si linkage by covering silicon surface.^26–28 Many other efforts were tried to add additives in electrolyte, such as FEC^29–31 and LiBOB,³² to form passivation layer on material to reduce side reactions.

As seen from the above, the surface properties of silicon particles are closely related to the SEI structure, electrodes stability and electrochemical cycling of Si-based anode. Previously, most studies have focused on the study of inherent surface oxide-layer on silicon particles, or tried to add a variety of additives in electrolyte solutions to form more stable passivation layer covering silicon particles. But few researchers have been devoted to surface modification by directly introducing new groups on silicon surface to optimize nano-particles and reduce side reactions with electrolyte solutions.

In this paper, three kinds of surface modifications on silicon nanoparticles were accomplished by the chemical treatments, corresponding to the samples of silanol silicon (Sis), carboxyl silicon (Sic) and siloxane silicon (Sio), respectively.^33–35 We compared the SEI structures and electrochemical performance of these decorated Si anode materials, and provide the possible mechanism of SEI film formation during initial lithium alloying of Si-particles.

2. Experimental methods

2.1. Synthesis of materials

The used Si-anodes are commercial crystallized silicon powder (∼40 nm in diameter). To explore the influence of various groups on Si-particle surface during lithiation/delithiation, we selected three common and stable modification groups on particles surface. Three kinds of different chemical-bonds were prepared on particles surface (can be seen in Fig. 1): silanol silicon, carboxyl silicon and siloxane silicon, respectively. Silanol silicon: 1 g crystallized-Si powder (∼40 nm in diameter) was dissolved in concentrated hydrochloric acid (HCl) solution and refluxed 6 hours to activate groups on Si-particle surface. Then repeatedly centrifugal and drying by DI-water and ethanol, surface group was removed by HF solution to expose a fresh surface. The cleaned Si-powder was soaked in Piranha solution (H₂SO₄ : H₂O₂ = 4 : 1) at 85 °C to oxidize 30 minutes, the solution was filtered and the residue was repeatedly washed by DI-water, and finally the achieved Si nano-particles was dried by drying nitrogen. Carboxyl silicon was prepared from silanol silicon: the precursor particles (synthesized silanol silicon) were dispersed in ethanol solution, and then added 3-aminopropyltriethoxysilane (APTES) in solution and stirred for 4 hours at room temperature, and the obtained material was centrifuged repeatedly to obtain pure Si nanoparticles. Then these were dissolved in DMF solution which was added to the 0.2 g butanedioic anhydride (C₄H₄O₃) and stirred for 12 hours. The mixture was also filtered and the residue was repeatedly washed by DI-water, and finally the achieved carboxyl-Si was dried by drying nitrogen. Siloxane silicon: silanol-Si was heated at 500 °C for 4 hours on the air circumstance in order to make ether bond by one-step dehydroxylation.

Fig. 1 Schematic of the chemical modification on surface silicon.

2.2. Materials characterizations

FT-IR spectra were measured on a VERTEX-70 (Bruker, Germany) spectrometer in the Diffuse Reflectance mode, which were conducted over the scanning range of medium-infrared (400–4000 cm⁻¹) with a Harrick HVC accessory. Different groups on Si-particle surface, such as Si–O–Si, Si–OH, Si–COOH, are susceptible to different movement modes in the medium infrared range. X-ray photoelectron spectroscopy (XPS) was conducted by using a Thermo ESCALAB-MKII 250 spectrometer (UK) and a focused mono-chromatized Al Kα radiation (1486.6 eV). The residual pressure in-side the XPS analysis chamber was 9.3 × 10⁻⁷ Pa. To avoid any contamination from oxygen and water, the cycled electrodes were transferred from the Ar-filled glove box to the XPS chamber using a special air-proof device. And the raw spectra were curve-fitted by non-linear least squares fittings with a Gauss–Lorentz ratio (80 : 20) through the XPSPEAK41 software.

2.3. Electrode preparation and electrochemical measurements

The synthesized-Si material were dried under vacuum at 80 °C and mixed with Super P, poly(acrylic acid) (PAA) in a mass ratio of 8 : 1 : 1 in N-methyl-2-pyrrolidone, and the resulting slurry was coated on the copper foil (18 μm) by semi-automatic coating-machine in a dry room. The coated electrode was dried for 48 h at 80 °C under vacuum to remove solvent. Test electrodes 1.8 cm in diameter was cut from the coated electrodes; the loading of electrodes on the foil were about 0.1 mg. CR2032 coin-type cells were assembled in an argon-filled glove box with the content of oxygen and water both less than 0.5 ppm. The cell consisted of a metallic lithium foil (thickness 400 mm) and the prepared Si-based electrodes, separated by glass fiber paper (thickness 1 mm, pore size 1.6 mm), and filled with a solution of 1 M LiPF₆ in 3 : 7 ethylene carbonate (EC) and ethyl methyl carbonate (EMC) as the electrolyte. Cells were placed in electrochemically tested at room temperature under galvanostatic cycling applied using a potentiostat (LAND). Full discharge and charge testing was performed over the voltage range of 0.005–2 V with a current density of 200 mA g⁻¹. To test the cycling performance, charge/discharge cycling tests were carried out in the time-controlled mode at a current density of 200 mA g⁻¹ for 5 hours. The electrocatalytic performance of three Si electrodes were tested by cyclic voltammetry (CV) and Electrochemical impedance spectroscopy (EIS) in a conventional three-electrode cell containing a Si electrode (work electrode) and two lithium plate reference electrode (RE) and counter electrode (CE). Cyclic voltammetry (CV) measurements were carried out at room temperature using the electrochemistry workstation, scanning at the different rate in the voltage range of 0.005–2 V. EIS experiments were taken under open circuit voltage (OCV) with a current amplitude of 5 mV over frequency range of 1 MHz to 10 mHz. All the tests were carried out at room temperature. The changes of impedance were simulated by Zview software.

3. Results and discussion

3.1. Physicochemical properties of the materials

In order to verify the chemical composition of Si-particles surface, the FT-IR was used to detect the movement state of modified-groups on particles surface. The FT-IR spectra of different materials are shown in Fig. 2. It contains three kinds of modified-silicon, Sic, Sis and Sio, respectively. Fig. 2 shows a clear disparity at three kinds of material in the frequency regions characteristic of Si–O–H and Si–C=O and coupled with Si–X stretching. For Sis, two wide peaks at 3000–3800 cm⁻¹ of the Si–O–H region is produced because of the stretching movement of hydroxyl group on Si-particle surface, and it contains the Si–O–H_free and Si–O–H_bond stretching.^36,37 After heated, the most of silanol are became siloxane group by dehydration. From Fig. 2, the peak of Si–OH on Sio electrodes is weakly and unobviously, replaced by the double-peak at 2256 cm⁻¹ position, which corresponds to O₂–Si–H_x and O₃–Si–H_x stretching modes.³⁸ After hydroxyl group was substituted by carboxyl group on Si-particles surface, the C=O double bonds of carboxyl have a unique vibration peak at around 1705 cm⁻¹.³⁹ And, the stretching peak of –CO group at 1775 cm⁻¹ is an important characteristic peak.⁴⁰ At lower wavenumbers regions (inset of Fig. 2b), the tiny peaks are present at 700–1000 cm⁻¹, such as the Si–O–H vibration modes at 950 cm⁻¹, the O–Si–H₂ scissor mode at 880 cm⁻¹, and the peak of Si–C at 805 cm⁻¹.⁴¹ Finally, because of the unstable surface of Si-particles, it is inevitable to produce partly structural rearrangement on silicon surface, corresponding to the strong absorption peak of O–Si–O bonds at 1198 cm⁻¹ wavenumbers.^38,42

Fig. 2 Diffuse infrared reflectance spectra of the sample Sic (blue spectrum), Sis (red spectrum) and Sio (black spectrum).

Fig. 3 show the Transmission Electron Microscope (TEM) images of three kinds of modification silicon. As can be seen from the figure, the thickness of surface layer of silicon nanoparticles has been changed obviously owing to chemical modification. The thickness of Sis (Fig. 3a) and Sic (Fig. 3c) was exhibited thinner than the Sio (Fig. 3b). For siloxane silicon, the surface layer was oxidized after temperature dehydration. In addition, the surface of carboxyl silicon was changed to irregularity owing to the modification of carboxyl group.

Fig. 3 The Transmission Electron Microscope (TEM) of three kinds of modification silicon, (a) Sis, (b) Sio, (c) Sic.

3.2. Electrochemical characteristic

To study the electrochemistry performance of different surface on silicon, three electrochemical methods were used, including electrochemistry charge and discharge (ECD), cyclic voltammetry (CV) and cycling tests. Electrode plates were transferred in glovebox without exposure to air in a sealed vacuum transfer cube, where they were then assembled into coin cells. Fig. 4 shows the first electrochemical alloying/dealloying curves of three kinds of Si electrodes at current 200 mA g⁻¹. Different phenomena can be easily distinguished from these curves. At first alloying, the particle surface is reacted with the electrolyte at reduction potential from 1.5 to 0.5 V (versus Li/Li⁺). These irreversible reactions can consume lithium and electrolyte components to form SEI, which is the structure of a lithium-ion conductive but electrolyte-blocking interface layer. Upon the following process, an obvious plateau at potentials negative to 100 mV can be observed for the formation of Li_xSi alloying.⁴³ The first cycle discharge (alloying) capacities are 2907, 3141, and 3195 mA h g⁻¹, corresponding to Sis, Sio and Sic electrodes, respectively. During the process of dealloying, reversible Li_xSi are dealloyed from silicon particles at around 300 mV (versus Li/Li⁺). The charging (dealloying) capacity of three kinds of electrodes is quite different, and Sic is devoted to the highest capacity (∼2184.82 mA h g⁻¹) followed by Sio (∼1894.06 mA h g⁻¹) and Sis (∼1622.2 mA h g⁻¹). Then, the initial coulombic efficiencies (ICE) are 68.1%, 60.3% and 55.8% for three kinds of silicon electrodes, Sic, Sio, Sis, respectively. The Sic in three kinds of electrodes not only have the highest intercalation lithium capacity, also the ICE is also the highest, which is owing to the formation of SEI consuming the less lithium. But, for Sis electrodes, the initial electrochemical performance is obviously worse. It is well-known that the silicon anode exist an irreversible capacity during first charging–discharging. Now, most studies considered that the lithium cannot be completely extracted from the alloying state of Li_xSi_y owing to the too large diffusion path of lithium in silicon phase for larger particles. In this study, we choose raw material of nano-sized silicon particles and only surface modification for material. Then, the difference of electrochemistry properties can be considered as the impact of surface group for three kinds of silicon material.

Fig. 4 The initial capacity curves of the different sample Sic (blue spectrum), Sio (red spectrum) and Sis (black spectrum).

Fig. 5a shows the cycling behavior of three kinds of electrodes, which illustrates that Sic exhibits higher discharge capacity than the values of Sis and Sio at beginning stage. It reflects Sic material only consume fewer Li-ion to form SEI at alloying. After 150 cycles, the Sic capacity were decreased to 1345.1 mA h g⁻¹, in relative to Sio and Sis electrodes were kept the capacity of 1540.8, 737.3 mA h g⁻¹, respectively (Fig. 5b). And it explains the stability of SEI on Sio electrodes exceed to the Sic (retention ratio: 81.33% vs. 61.56%), and its result may be due to the SEI on Sic was gradually cracked after the repeat volume expansion, but the Sio's SEI was excellent. Owing to the crack of SEI, it is lead to the exposed of fresh surface on silicon particles and rapid exhausted lithium in electrolyte. Moreover, Sis electrodes have lower reversible capacity the same as the initial capacity curves from Fig. 4. For Sio, although it has lower specific capacity, the cycling capacity is stability, which indicates the excellent formation of SEI. Fig. 5c exhibits the discharge and rate performance of three kinds of silicon. As can be seen, the rate capability of Sic was shown the same as Sio silicon, but the capacity of Sis was rapidly fading.

Fig. 5 The cycling curves of discharge capacity of three kinds of silicon electrode, (a) discharge capacity and coulomb efficiency, (b) schematic of capacity retention and (c) rate capability.

Fig. 6 depicts the cyclic voltammograms (CV) of three kinds of electrodes measured over range of 2–0.005 V at a scan rate of 0.1, 0.2, 0.4, 0.8 mV s⁻¹. And at the 0.1 mV s⁻¹, electrodes were carried out for three scans repeatedly. It can be found that there are similar curves at the three scans at 0.1 mV s⁻¹ and the subsequent cycles of different scan rate. In Fig. 6, two obvious current peaks appear at around 0.6 V and 0.25 V in the first cathodic scanning, while only those below 0.25 V which can be attributed to the reversible alloying. The broad reaction peak around 0.6 V was considered the formation of SEI and side-reactions at the interface. It knows that the structure and stability of SEI have an important influence on electrochemical properties. The peak below 0.25 V is attributed to the formation of Li_xSi_y alloy, which be associated with the reversible alloying/dealloying reaction with Li-ion.^3,44,45 During the second and third scanning on 0.1 mV s⁻¹, the cathodic curves of Sic (Fig. 6a) and Sio (Fig. 6b) become almost stable, which are indicative of the reversible behavior of lithiation/delithiation. But for Sis, the current of first anodic and subsequent scan was decreased after the first cathodic scanning. At the subsequent cyclic voltammetry of different scanning rate (∼0.2, 0.4, 0.8 mV s⁻¹), the cathodic and anodic currents of Sic and Si were gradually increased and became obviously, but the peak of Sis was irreversible.

Fig. 6 CV plots at different scanning rate for (a) Sic, (b) Sio, (c) Sis.

3.3. Analysis of cycling stability

The results of discharge/charge, cyclic and CV curves show that different modified-groups on surface plays a significant role in determining the electrochemical properties, thus, it is necessary to investigate what are the SEI ingredients on these electrodes. The XPS spectra provide insights about the formation of SEI components. For this purpose, three kinds of electrodes were characterized using the high-resolution XPS techniques at end of the first alloying. Fig. 7 shows C 1s, O 1s, F 1s, Si 2p XPS spectra of three modified-Si electrodes obtained with middle energy (∼1486.6 eV), and Fig. 7a–l corresponds to Sis, Sic and Sio electrodes, respectively. The spectra of C 1s display several components, with the most intensive peak at around 284.2 eV (blue in Fig. 7a, e and i) corresponds to hydrocarbon species and carbon black, the peak at around 285.5 eV (green) can be attributed to single C–O bonds, a middle peak around 288.1 eV (red) can resulted in alkoxy carbon (O=C–O bonds), the intensive peak at around 289.8 eV (brown) is typical of carbonate groups (CO₃ bonds).^37,46 The overall shape of C 1s spectra is similar upon three kinds of electrodes. Nevertheless, the SEI of Sis electrodes contains more components of carbonate lithium and less alkoxy carbon and C–O bonds, which is dependent on the decomposition from solvent in electrolyte (ref. reactions in Scheme 1).¹¹ But for Sic and Sio electrodes, the peak intensity of three components (–CO₃, O=C–O, C–O) in spectra are almost equal.

Fig. 7 C 1s, O 1s, F 1s and Si 2p spectra of (a–d) Sis, (e–h) Sic, (i–l) Sio-based electrodes at the alloying.

Scheme 1 Reduction reaction of EC-EMC/LiPF₆.

Fig. 7b, f and j are presented the O 1s spectra of three kinds of electrodes at the alloying. The O 1s spectra shows a broad peaks at 530–536 eV that can be considered into three individual peaks at 531, 532.8, 534.6 eV, which be related to the O 1s binding energies of Li_xSiO_y, Li₂CO₃ and polycarbonate species, respectively.⁴⁷ And, the peak at 534.6 eV in spectra should be attributed to a contribution from ROCO₂Li (mainly species, resulting from EC and EMC reduction with surface groups, Scheme 1). From Fig. 7b, we can see that Sis electrodes have an abundance of lithium carbonate which is consistent with the conclusion of the C 1s spectra. However, Sic and Sio electrodes form the SEI of rich-ROCO₂Li whereas the peak intensity of Li_xSiO_y at 534.6 eV is also high in Sic electrodes.

It is well known that electrolytic (the main is LiPF₆) are decomposed with trace water on electrodes at the first intercalation, and LiF is one of the important species in these side-reaction production. Thus, the F 1s spectra are also conducted to obtain LiF concentration in the different SEI structures. The spectra (Fig. 7c, g and k) shows two peaks at around 685.2, 687.2 eV, the first peak is contributed to lithium fluriode (LiF) of the degradation products of LiPF₆. Another peak (∼687 eV) is due to the salt LiPF₆ remaining at the surface of electrodes.²⁶ Sis electrodes were measured with higher relative intensity of LiF (Fig. 7c), which reflects more side-reactions than another electrodes to consume the lithium salt (LiPF₆) in electrolyte. And a relatively low intensity of LiPF₆ peak (∼687 eV) was again confirmed the decomposition of Li salt. But for Sic, the intensity of F 1s spectra (Fig. 7g) is the opposite, containing the strong peak of LiPF₆ and the weak of LiF, which indicates that the less exhaustion of lithium can keep higher capacity than others at first alloying. Comparing the spectra of Sio and Sic electrodes, the Sio's SEI components contain more LiF (∼685 eV in F 1s spectra) and Li₂CO₃ species (∼534.6 eV in O 1s spectra), the results are shown as lower ICE but more stable cycling performance.

The Si 2p spectra contain a broad peak at 101.5 and 103.5 eV consistent with the presence of Li_xSiO_y and SiO₂ species, and another peak for Li_xSi (∼98 eV).^26,48 In spectra of Fig. 7d, the Li_xSi peak of Sis was not observed, it can be seen the SEI thickness of surface exceed to the analysis depth for XPS. And the Sic and Sio present the strong intensity peak of Li_xSi and Li_xSiO_y. And, the Sio electrodes has a high intensity of Li_xSi peak (∼98 eV) indicative of the thin SEI layer than Sic.

Combining the XPS information provided by the C 1s, O 1s, F 1s and Si 2p spectra reveal that Sio electrodes forms uniform SEI with high concentration of LiF and ROCO₂Li at the first alloying, and in relative to the SEI of Sic including more species of ROCO₂Li, Li_xSiO_y. But for Sis electrodes, the lithium salt (∼LiPF₆) in electrolyte was degraded to form side-reaction of Li₂CO₃ and LiF on particles surface.

3.4. Analysis of EIS

To further explore the reasons of different performance of cycling, EIS measurement was conducted at different cycles, 1, 5, 10, 50 cycles, respectively. The Nyquist plots and corresponding equivalent circuit diagrams are presented in Fig. 8, which illustrates that the charge-transfer resistance was influenced by different cycles. The EIS curves contain the semicircle in high-frequency and a following line in low-frequency. The intercept of super-high frequency at real part (Z′) corresponds to the equivalent series resistance (R_esr), which is mainly induced from the ohmic resistance of the electrodes. The diameter of the semicircle in high-frequency represents the charge-transfer resistance (R_ct) containing the faradic reactions and the double-layer capacitance (C_dl) between electrode and electrolyte.⁴⁹ A slop line in the low-frequency range is called the Warburg resistance (Z_w) and is relative to the ionic transport and diffusion resistance.^50,51

Fig. 8 EIS performance at different cycles (1, 5, 10, 50 cycles, respectively). (a) Nyquist plots of Sic. (b) Equivalent circuit diagrams. (c) Nyquist plots of Sis, and (d) Sio.

As shown in Fig. 8, we can see the continuous increasing of R_ct in high-frequency (∼semicircle diameter) during cycles, and the reasons can be considered as a result of the changes of SEI of electrodes. In Fig. 8c, Sis demonstrate higher resistance of high frequency than Sic and Sio electrodes at first cycle, and the transfer resistance also rapidly increase and resulting in a worse interface during cycling. Moreover, it is indicating, at first alloying, the Sis did not form the stability SEI, and the hydroxyl-group on surface was involved in the occurrence of side-reaction. Then, the interface of Si electrode is gradually deteriorated to lead to poor contact and electrolyte consumption. In Fig. 8c, the Sic demonstrates exhibits a lowest R_ct (∼70.9 ohm cm) values compared with those of Sio (∼73.6 ohm cm) and Sis (∼90.6 ohm cm) with the result of the thinner SEI and the higher ICE. But after cycling, the increasing of resistance is caused by the instability of SEI on undergoing volume expansion of Si-particles. For Sio, an oxide layer on surface is easy to react with electrolyte to produce a homogeneous SEI. The SEI layer mainly includes lithium fluoride and lithium polycarbonates, which can resist the volume chang of particles and ensure the excellent cycling performance and resistance properties. Through the Sio demonstrates the fastest charge transfer rate which would be contributable to the best cycling capacity.

3.5. Mechanism

According to the previous discussion, the result of XPS and EIS reflect the difference between the SEI's structures on surface of silicon nanoparticles. And the different components of SEI take an obvious influence on the electrochemical performance though CV and cycling. The mechanism of the formation of SEI on different surface of Si-particles is not yet clear, but we also provides a schematic to express the possible mechanism on observed results. Fig. 9 summarizes the formation process of SEI at the first alloying for three kinds of Si-particles surface. The Sis (Fig. 9a), owing to the presence of abundant hydroxyl groups, are reacted with electrolyte at the low reduction potential and exhausted a large number of Li-ion lead to the low initial irreversible efficiency. Moreover, due to the high concentration LiF species in SEI, we can think that the hydroxyl group on surface are formed the H₂O molecular at potential and further decomposed with the LiPF₆ of electrolyte. And for Sic (Fig. 9b) and Sio (Fig. 9c), both electrodes can form excellent SEI, which contains component of Li_xSiO_y, ROCO₂Li and LiF, Li₂CO₃, etc., respectively. But the transfer resistance of SEI for Sic was gradually increased during cycling, and the instability of SEI can be difficult to resist the volume expansion of Si-particles.

Fig. 9 Schematic view of the mechanism occurring on the surface of three kinds of silicon particles. Formation of the SEI at the beginning of alloying for (a) Sis, (b) Sic and (c) Sio.

4. Conclusion

Owing to nanosized Si-particles inevitably has an oxide layer (or silica layer) on surface, the silica layer plays a vital role during the charge/discharge, which could react with electrolyte at reduction potential to generate a variety of lithium salt, such as lithium carbonate, lithium fluoride, lithium silicate and other complex compounds. And during cycling, the volume changes of Si-particles could lead to the fracture of particles and expose fresh surface to electrolyte so that the side-reaction was occurred on surface of electrodes. In this paper, we have synthesized three kinds of Si-based material with different surface group, containing silanol silicon, carboxyl silicon, siloxane silicon. The electrochemical properties of three kinds of Si-based materials were compared through charge/discharge, cyclic test and cyclic voltammetry. Through XPS, we knows that different components of SEI for three kinds of silicon material. Through analysis of EIS, it exhibits that the hydroxyl group on surface of Sis was decomposed with electrolyte and cycling capacity was declined (retention: 50% at 150 cycles), and Sio can formed the excellent SEI and cycling performance. For Sic, the formation of thinner SEI can provides the high initial capacity (2184.82 mA h g⁻¹) but the lack of stability results in the capacity fading (retention: 61.56% at 150 cycles) by repeatedly volume expansion.

Acknowledgements

This work was financially supported by National Natural Science Foundation of China (21571173) and Science & Technology Office of Jiangsu Province (BE2013006-3).

References

1. C. K. Chan, R. Ruffo, S. S. Hong and Y. Cui, J. Power Sources, 2009, 189, 1132.
2. D. E. Arreaga-Salas, A. K. Sra, K. Roodenko, Y. J. Chabal and C. L. Hinkle, J. Phys. Chem. C, 2012, 116, 9072.
3. M. N. Obrovac and L. J. Krause, J. Electrochem. Soc., 2007, 154, A103.
4. B. P. N. Nguyen, J. Gaubicher and B. Lestriez, Electrochim. Acta, 2014, 120, 319.
5. U. Kasavajjula, C. Wang and A. J. Appleby, J. Power Sources, 2007, 163, 1003.
6. H. Kim, B. Han, J. Choo and J. Cho, Angew. Chem., Int. Ed., 2008, 120, 10305.
7. A. Magasinski, P. Dixon, B. Hertzberg, A. Kvit, J. Ayala and G. Yushin, Nat. Mater., 2010, 9, 353.
8. R. Z. Hu, W. Sun, Y. L. Chen, M. Q. Zeng and M. Zhu, J. Mater. Chem. A, 2014, 2, 9118.
9. M. Sun, R. Z. Hu, H. Liu, M. Q. Zeng, L. C. Yang, H. H. Wang and M. Zhu, J. Power Sources, 2014, 268, 610.
10. X. H. Liu, L. Zhong, S. Huang, S. X. Mao, T. Zhu and J. Y. Huang, ACS Nano, 2012, 6, 1522.
11. X. Su, Q. Wu, J. Li, X. Xiao, A. Lott, W. Lu, B. W. Sheldon and J. Wu, Adv. Energy Mater., 2014, 4, 1300882.
12. D. Aurbach, J. Power Sources, 2000, 89, 206.
13. K. Schroder, J. Alvarado, T. A. Yersak, J. Li, N. Dudney, L. J. Webb, Y. S. Meng and K. J. Stevenson, Chem. Mater., 2015, 27(16), 5531.
14. A. M. Andersson, A. Henningson, H. Siegbahn, U. Jansson and K. Edström, J. Power Sources, 2003, 119–121, 522.
15. D. Aurbach, B. Markovsky, M. D. Levi, A. Schechter, M. Moshkovich and Y. Kohen, J. Power Sources, 1999, 95, 81.
16. K. C. Moller, T. Hodal, W. K. Appel, M. Winter and J. O. Besenhard, J. Power Sources, 2001, 595, 97.
17. J. Christensen and J. Newman, J. Electrochem. Soc., 2005, 152, A818.
18. R. Deshpande, M. W. Verbrugge, Y. T. Cheng, J. Wang and P. Liu, J. Electrochem. Soc., 2012, 159, A1730.
19. D. D. D. Ma, C. S. Lee, F. C. K. Au, S. Y. Tong and S. T. Lee, Science, 2003, 299, 1874.
20. B. C. Yu, Y. Hwa, C. M. Park, J. H. Kim and H. J. Sohn, RSC Adv., 2013, 3, 9408.
21. Y. S. Hu, R. Demir-Cakan, M. M. Titirci, J. O. Muller, R. Schlogl, M. Antonietti and J. Maier, Angew. Chem., Int. Ed., 2008, 47, 1645.
22. K. Xu and A. Von Cresce, J. Mater. Chem., 2011, 21, 9849.
23. B. Philippe, R. Dedryvère, J. Allouche, F. Lindgren, M. Gorgoi, H. Rensmo, D. Gonbeau and K. Edström, Chem. Mater., 2012, 24, 1107.
24. S. Xun, X. Song, L. Wang, M. E. Grass, Z. Liu, V. S. Battaglia and G. Liu, J. Electrochem. Soc., 2011, 158(12), A1260.
25. S. W. Song and S. W. Baek, Electrochem. Solid-State Lett., 2009, 12(2), A23.
26. Y. G. Ryu, S. S. Lee, S. K. Mah, D. J. Lee, K. Kwon, S. S. Hwang and S. G. Doo, J. Electrochem. Soc., 2008, 155, A583.
27. S.-W. Song and S.-W. Baek, Electrochem. Solid-State Lett., 2009, 12(2), A23.
28. G. Schroeder, B. Gierczyk, D. Waszak and M. Walkowiak, Electrochem. Commun., 2006, 8, 523.
29. V. Etacheri, O. Haik, Y. Goffer, G. A. Roberts, I. C. Stefan, R. Fasching and D. Aurbach, Langmuir, 2012, 28, 965.
30. A. Bordes, K. Eom and T. F. Fuller, J. Power Sources, 2014, 257, 163.
31. Y. M. Lin, K. C. Klavetter, P. R. Abel, N. C. Davy, J. L. Snider, A. Heller and C. B. Mullins, Chem. Commun., 2012, 48, 7268.
32. N. S. Choi, K. H. Yew, H. Kim, S. S. Kim and W. U. Choi, J. Power Sources, 2007, 172, 404.
33. J. Gustavsson, G. Altankov, A. Errachid, J. Samitier, J. A. Planell and E. Engel, J. Mater. Sci.: Mater. Med., 2008, 19, 1839.
34. W. J. Xu, J. Rytkönen, S. Rönkkö, T. Nissinen, T. Kinnunen, M. Suvanto, A. Närvänen and V.-P. Lehto, Chem. Mater., 2014, 26, 6734.
35. K. Jradi, D. Laour, C. Daneault and B. Chabot, Colloids Surf., A, 2011, 374, 33.
36. J. P. Gallas and J. C. Lavalley, Langmuir, 1991, 7, 1235.
37. U. S. Vogl, P. K. Das, A. Z. Weber, M. Winter, R. Kostecki and S. F. Lux, Langmuir, 2014, 30, 10299.
38. B. Douglas, J. A. Mawhinney, J. T. Glass and J. Yates, J. Phys. Chem. B, 1997, 101, 1202.
39. B. J. Koo, H. J. Kim, Y. Y. Cho, K. T. Lee, N. S. Choi and J. Cho, Angew. Chem., Int. Ed., 2012, 124, 8892.
40. Z. J. Han, N. Yabuuchi, S. Hashimoto, T. Sasaki and S. Komaba, ECS Electrochem. Lett., 2013, 2(2), A17.
41. A. Burneau and O. Barrès, Langmuir, 1990, 6, 1364.
42. J. H. Song and M. J. Sailor, Inorg. Chem., 1999, 38, 1498.
43. T. Zhang, J. Gao, H. P. Zhang, L. C. Yang, Y. P. Wu and H. Q. Wu, Electrochem. Commun., 2007, 9, 886.
44. Z. Li, X. J. Huang, L. Q. Chen, G. W. Zhou, Z. Zhang, D. P. Yu, Y. J. Mo and N. Pei, Solid State Ionics, 2000, 135, 181.
45. J. G. Tu, Y. Yuan, P. Zhan, H. D. Jiao, X. D. Wang, H. M. Zhu and S. Q. Jiao, J. Phys. Chem. C, 2014, 118, 7357.
46. S. P. V. Nadimpalli, V. A. Sethuraman, S. Dalavi, B. Lucht, M. J. Chon, V. B. Shenoy and P. R. Guduru, J. Power Sources, 2012, 215, 145.
47. B. Philippe, R. Dedryvère, J. Allouche, F. Lindgren, M. Gorgoi, H. Rensmo, D. Gonbeau and K. Edström, Chem. Mater., 2012, 24, 1107.
48. E. Radvanyi, W. Porcher, E. D. Vito, A. Montani, S. Franger and S. J. S. Larbi, Phys. Chem. Chem. Phys., 2014, 16, 17142.
49. A. Barai, G. Chouchelamane, Y. Guo, A. McGordon and P. Jennings, J. Power Sources, 2015, 280, 74.
50. H. M. Cho, W. S. Choi, J. Y. Go, S. E. Bae and H. C. Shin, J. Power Sources, 2012, 198, 273.
51. L. Liao, P. Zuo, Y. Ma, X. Chen, Y. An, Y. Gao and G. Yin, Electrochim. Acta, 2012, 60, 269.

---