Research Article www.acsami.org
Constructing Novel Si@SnO2 Core−Shell Heterostructures by Facile Self-Assembly of SnO2 Nanowires on Silicon Hollow Nanospheres for Large, Reversible Lithium Storage Zheng-Wei Zhou,† Yi-Tao Liu,*,†,‡ Xu-Ming Xie,*,† and Xiong-Ying Ye‡ †
Key Laboratory of Advanced Materials (MOE), Department of Chemical Engineering, Tsinghua University, Beijing 100084, China State Key Laboratory of Precision Measurement Technology and Instruments, Department of Precision Instrument, Tsinghua University, Beijing 100084, China
‡
S Supporting Information *
ABSTRACT: Developing an industrially viable silicon anode, featured by the highest theoretical capacity (4200 mA h g−1) among common electrode materials, is still a huge challenge because of its large volume expansion during repeated lithiation− delithiation as well as low intrinsic conductivity. Here, we expect to address these inherent deficiencies simultaneously with an interesting hybridization design. A facile self-assembly approach is proposed to decorate silicon hollow nanospheres with SnO2 nanowires. The two building blocks, hand in hand, play a wonderful duet by bridging their appealing functionalities in a complementary way: (1) The silicon hollow nanospheres, in addition to the major role as a superior capacity contributor, also act as a host material (core) to partially accommodate the volume expansion, thus alleviating the capacity fading by providing abundant hollow interiors, void spaces, and surface areas. (2) The SnO2 nanowires serve as a conductive coating (shell) to enable efficient electron transport due to a relatively high conductivity, thereby improving the cyclability of silicon. Compared to other conductive dopants, the SnO2 nanowires with a high theoretical capacity (790 mA h g−1) can contribute outstanding electrochemical reaction kinetics, further adding value to the ultimate electrochemical performances. The resulting novel Si@SnO2 core−shell heterostructures exhibit remarkable synergy in large, reversible lithium storage, delivering a reversible capacity as high as 1869 mA h g−1@500 mA g−1 after 100 charging−discharging cycles. KEYWORDS: core−shell heterostructures, lithium-ion batteries, self-assembly, silicon, SnO2
■
INTRODUCTION Today, human’s urgent demand for next-generation lithium-ion batteries (LIBs) with higher energy densities and rate capabilities, capable of driving large-size power tools such as electrified vehicles, has provoked worldwide interest in developing high-capacity electrode materials.1 In this context, silicon stands out as a prospective anode candidate due to its rich resources, low cost, and high safety. More importantly, silicon outperforms other common electrode materials in terms of theoretical capacity (4200 mA h g−1) because of its ability to accommodate 4.4 mol of lithium per mole of Si (Li22Si5),2−5 which is more than 10 times that (372 mA h g−1) of the current LIB anode, i.e., graphitic carbon. Despite these merits, however, developing an industrially viable anode exclusively composed of silicon is still a huge challenge not yet realized now, mainly due to its large volume expansion (∼400%) during repeated lithiation−delithiation as well as low intrinsic conductivity (∼10−3 S cm−1). On the one hand, the large volume expansion is responsible for the serious pulverization of the silicon anode, which causes electrical disconnection from the current collector © XXXX American Chemical Society
and parasitic formation of a solid−electrolyte interface (SEI) layer on the freshly exposed silicon surfaces. On the other hand, the low intrinsic conductivity unfortunately leads to a high cell resistance, further hastening the electrolyte degradation. The two inherent deficiencies of silicon, as the chief reasons for rapid capacity fading especially when cycled at high rates, remain to be addressed before its successful implementation in next-generation, high-power LIBs. Basically, the deficiency of large volume expansion can be partially addressed by engineering silicon into various nanostructures, e.g., hollow nanospheres,6,7 nanosheets,8 nanowires,9−13 or nanoparticles,14−16 which are expected to improve the mechanical integrity and minimize the volume change during the charging−discharging processes. It is worth noting, however, that the extremely large surface areas of these nanostructures inevitably induce their aggregation upon cycling, Received: January 5, 2016 Accepted: February 29, 2016
A
DOI: 10.1021/acsami.6b00107 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
lithium storage, delivering a reversible capacity as high as 1869 mA h g−1@500 mA g−1 after 100 charging−discharging cycles. Note that our self-assembly approach for the hybridization design is really facile and energy-efficient compared to those complicated techniques employing nanosphere lithography combined with inductively coupled plasma (ICP) dry etching plus radio frequency (RF) magnetron sputtering48 and is therefore more promising in practical LIB processing.
resulting in an obvious loss of active sites for lithium storage. Moreover, the deficiency of low intrinsic conductivity that also deteriorates the cyclability of the silicon anode is still a dilemma waiting for an efficient resolution. To enhance the electrochemical performances of the silicon anode, another effort is taken to encapsulate it with a flexible conductive coating such as amorphous carbon.17−22 However, the synthesis of the carbon coating, mainly through thermal decomposition of a carbon precursor at high temperatures, is usually accompanied by the release of volatile organic compounds that may cause unbearable environmental problems. More recently, some hard-confining conductive matrices, for instance, carbon nanotubes,23−26 carbon nanofibers,27 and reduced graphene oxide (r-GO),28−33 are employed to host various silicon nanostructures, thereby increasing their electrical conductivity. It should be emphasized, however, that the cycle and rate performances of a carbon-doped silicon anode are largely affected by the quality of the carbon dopant, which is determined by its preparation method and is often inadequate and nonuniform. For example, r-GO, although advocated as a high-efficiency matrix, is structurally permanently damaged during the oxidation−reduction process, and its electrical conductivity is thus significantly destroyed compared to graphitic carbon.34 Besides, the consumption of large quantities of toxic oxidants and/or reductants for producing r-GO also imposes intense environmental concerns. In this sense, more effective conductive dopants such as noble metals (Au,35,36 Ag,37−40 Ge,41−43 etc.) are emerging to facilitate electron transport due to their excellent conductivity resilience. However, the unaffordable cost of noble metals makes their conductive doping for silicon an armchair strategy unrealistic for commercialization now. Given the above analysis of existing circumstances for the silicon anode, here, we expect to address the two inherent deficiencies of silicon, i.e., large volume expansion and low intrinsic conductivity, simultaneously with an interesting hybridization design. A facile self-assembly approach based on van der Waals interactions, arising inherently from correlated electronic fluctuations in matter, is proposed to decorate silicon hollow nanospheres with SnO2 nanowires. The two building blocks, hand in hand, play a wonderful duet by bridging their appealing functionalities in a complementary way: (1) The silicon hollow nanospheres, in addition to the major role as a superior capacity contributor, also act as a host material (core) to partially accommodate the large volume expansion, thus alleviating the capacity fading by providing abundant hollow interiors, void spaces, and surface areas.44 (2) The SnO2 nanowires serve as a conductive coating (shell) to enable efficient, continuous one-way electron transport due to a relatively high electrical conductivity (up to 74 S cm−1),45 thereby improving the cyclability of silicon. It is worth stressing that, compared to other conductive dopants (e.g., nanocarbons and noble metals), the SnO2 nanowires featured by a high theoretical capacity (790 mA h g−1) can contribute outstanding electrochemical reaction kinetics, further adding value to the ultimate electrochemical performances.46,47 Besides, the advantages of the SnO2 nanowires over nanocarbons and noble metals also lie in their easy and green synthesis, low cost, high safety, and negligible environmental harm. Last but not least, the SnO2 nanowires can effectively isolate the silicon hollow nanospheres to ensure large void spaces and surface areas for lithium diffusion. The resulting novel Si@SnO2 core−shell heterostructures exhibit remarkable synergy in large, reversible
■
EXPERIMENTAL SECTION
Synthesis of Silicon Hollow Nanospheres. Monodisperse silica hollow nanospheres were synthesized by a hydrothermal reaction.49 Typically, vinyltriethoxysilane (2.5 g) was added to deionized water (50 mL) and stirred for 30 min to obtain a transparent solution, to which 25% NH3·H2O (0.5 mL) was added quickly. The obtained dispersion, after a 1 h reaction, was transferred to a Teflon-lined stainless steel autoclave, heated to 200 °C, and kept at that temperature for 24 h. The resulting silica hollow nanospheres were collected by centrifugation, washed by deionized water, and vacuumdried at 60 °C overnight. Next, the as-synthesized silica hollow nanospheres and magnesium powders at a molar ratio of 1:2.5 were placed at the opposite sides of a corundum boat and heated in a tube furnace at 700 °C for 5 h in an argon atmosphere containing 5% (v/v) hydrogen. The obtained brown powders, namely, monodisperse silicon hollow nanospheres, were washed by HCl and HF in sequence and vacuum-dried at 60 °C overnight.50 Synthesis of SnO2 Nanowires. As reported by us previously,46 SnCl4·5H2O (5 mmol), oleylamine (10 mmol), and oleic acid (40 mmol) were added to absolute ethanol (100 mmol) in a glass cup and stirred for 10 min. The glass cup was then transferred to a Teflon-lined stainless steel autoclave containing 20 mL of ethanol solution (96% v/ v). The autoclave was heated to 180 °C and kept at that temperature for 18 h. The obtained SnO2 nanowires were washed by absolute ethanol and vacuum-dried at 60 °C overnight. Construction of Si@SnO2 Core−Shell Heterostructures. In a typical experiment, the as-synthesized silicon hollow nanospheres and SnO2 nanowires were mixed in tetrahydrofuran (THF) at different weight ratios (90:10, 80:20, and 70:30). The mixtures were subjected to sonication (240 W) at ambient temperature for 24 h such that the SnO2 nanowires spontaneously assembled on the naked surfaces of the silicon hollow nanospheres through van der Waals interactions, thus forming Si@SnO2 core−shell heterostructures. The Si@SnO2 core− shell heterostructures were collected by vacuum filtration, followed by THF washing and vacuum-drying at 80 °C for 2 h. Evaluation of Electrochemical Performances. A working electrode was prepared by mixing 60 wt % active material, 20 wt % super P, and 20 wt % sodium alginate in deionized water, which was then coated on a copper foil.51 After being dried in a vacuum oven at 80 °C for 4 h, the working electrode was cut into a ∼1.3 cm2 circular sheet. A CR2025 coin cell was assembled in an argon-filled glovebox by using metallic lithium as the counter/reference electrode, 1 M solution of LiPF6 in ethylene carbonate/dimethyl carbonate (volume ratio = 1:1) as the electrolyte, and Celgard 2400 polypropylene as the separator. The galvanostatic charging−discharging experiment was performed by a battery tester LAND-CT2001A in a potential range of 0.01−1 V at ambient temperature. The mass loading was ∼1.2 mg cm−2 for the working electrode. The specific capacity was calculated on the basis of the whole active material. The CV measurement was conducted by a CHI600E electrochemical workstation at a scanning rate of 0.1 mV s−1 in a potential range of 0.01−1 V. The EIS measurement was also performed by this workstation. Characterizations. TEM was performed by a Hitachi HT7700 microscope operated at an accelerating voltage of 100 kV. HRTEM was performed by a JEOL JEM-2010 microscope operated at an accelerating voltage of 120 kV. SEM was performed by a Hitachi SU8010 microscope operated at an accelerating voltage of 5.0 kV. XRD was performed by a D8 Advanced X-ray diffractometer with Cu Kα radiation (λ = 0.154 nm). XPS was performed by an Escalab 250 Xi spectrometer. Raman was performed by a RamanMicro 300 B
DOI: 10.1021/acsami.6b00107 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces spectrometer (PerkinElmer, USA) with a 514 nm incident laser. BET measurement was performed by a Quantachrome Autosorb-1 Canalyzer (USA).
strong van der Waals forces and therefore facilely assemble on the naked surfaces of the silicon hollow nanospheres, forming micelle-like, stable Si@SnO2 core−shell heterostructures. In this way, the total free energy of the whole solution system is minimized by the formation of the SnO2 coating (shell). Note that the self-assembly based on van der Waals interactions is a newly developed, versatile strategy applicable to a number of 2D crystals including graphene and inorganic graphene analogues,52−56 all of which are featured by large surface areas and high surface energies. We stress that our hybridization design based on van der Waals interactions is really elegant compared to those complicated techniques employing nanosphere lithography combined with ICP dry etching plus RF magnetron sputtering,48 since the experimental conditions are mild and the energy consumption is low. The resulting Si@ SnO2 core−shell heterostructures are expected to facilitate lithium and electron transport and exhibit remarkable synergy in large, reversible lithium storage as will be proven later. The silicon hollow nanospheres can be converted from hydrothermally synthesized silica hollow nanospheres by magnesiothermic reduction.57 A morphological comparison between the two kinds of hollow nanospheres is realized by TEM characterization, as shown in Figure 2. As seen from this figure, the monodisperse silica hollow nanospheres, with obvious hollow interiors, have an outer radius (Rout) of around 215 nm and an inner radius (Rin) of around 190 nm (Figure 2a), corresponding to a wall thickness of ∼25 nm. The formation of the silica hollow nanospheres originates from an Ostwald ripening process, during which smaller pores are initially formed inside the nanoparticles, and the higher surface energy of these smaller pores makes them collapse into larger voids. These larger voids eventually merge and form hollow interiors. Meanwhile, a series of Si−O−Si bond-breaking inside and bond-making outside leads to the formation of the silica hollow nanospheres.49 Note that the silica hollow nanospheres can be adopted as a silicon source to prepare silicon hollow nanospheres by magnesiothermic reduction that has been well
■
RESULTS AND DISCUSSION Characterizations of Si@SnO2 Core−Shell Heterostructures. The whole process for constructing the Si@ SnO2 core−shell heterostructures from the two building blocks, i.e., silicon hollow nanospheres and SnO2 nanowires, is depicted in Figure 1. As mentioned above, the silicon hollow
Figure 1. As-synthesized silicon hollow nanospheres and SnO2 nanowires are mixed in THF under sonication, during which the SnO2 nanowires spontaneously assemble on the naked surfaces of the silicon hollow nanospheres through van der Waals interactions, thus forming Si@SnO2 core−shell heterostructures that can facilitate lithium and electron transport by bridging the appealing functionalities of the two building blocks in a complementary way.
nanospheres have large surface areas (237.5 m2 g−1, Figure S1) which translate into extremely high surface energies and strong van der Waals forces. Therefore, once low-surface-energy entities, e.g., the organically modified SnO2 nanowires in this case, are introduced, they will be spontaneously attracted by the
Figure 2. (a) TEM image of silica hollow nanospheres, (b−d) TEM and HRTEM images of silicon hollow nanospheres, and (e) XRD patterns of silica and silicon hollow nanospheres. The inset in (b) shows the ED pattern of silicon hollow nanospheres. C
DOI: 10.1021/acsami.6b00107 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces studied.57 Briefly, at a temperature above the melting point (650 °C), the magnesium powders are melted and flow into the mesopores of the silica hollow nanospheres, where the two react to form a MgO−Si nanocomposite according to the following formula: 2Mg(g) + SiO2 (s) → 2MgO(s) + Si(s)
(1)
After being thoroughly washed by HCl and HF, MgO is removed and silicon hollow nanospheres are left. It should be emphasized that the obtained silicon hollow nanospheres can preserve the original morphology and size without obvious deformation or collapse (Figure 2b). The inset shows the corresponding ED pattern of the silicon hollow nanospheres, which reveals (111), (220), and (311) diffraction rings of polycrystalline silicon.57 An enlarged TEM image focusing on an individual silicon hollow nanosphere unambiguously discloses a mesoporous structure with uniformly distributed primary silicon crystallites of 20−30 nm (Figure 2c).50 Under HRTEM observation, a measured lattice fringe spacing of ∼0.32 nm points to the (111) plane of polycrystalline silicon (Figure 2d). A further proof of the phase evolution from silica to silicon is provided by their XRD patterns, as shown in Figure 2e. No crystallization peaks are observed in the XRD pattern of the silica hollow nanospheres except for a broad one centered at 2θ = ∼22.0°, which is characteristic of amorphous silica (JCPDS card No. 29-0085).58 As to the XRD pattern of the silicon hollow nanospheres, the arising (111), (220), (311), (400), and (331) crystallization peaks correspond to highly crystalline silicon (JCPDS card No. 27-1402).59 A lattice fringe spacing of 0.31 nm is calculated according to Bragg’s equation based on the (111) crystallization peak located at 2θ = 28.3°, which agrees well with what is observed in the HRTEM image (Figure 2d). The abundant hollow interiors, void spaces, mesopores, and surface areas are expected to ensure short lithium diffusion pathways, accommodate volume changes, and allow for efficient strain relaxation. An examination on the overall morphology of the assynthesized SnO2 nanowires is achieved by TEM characterization, as shown in Figure 3a. It can be seen that the organically modified SnO2 nanowires are uniformly dispersed in THF at a high concentration, which can be attributed to the presence of a surface organic layer that renders the SnO2 nanowires highly oleophilic.60 The lengths of the SnO2 nanowires range from 70 to 100 nm, and the diameters have a narrow distribution within 4.1 ± 0.5 nm (85%).46 The corresponding ED pattern of the SnO2 nanowires is presented in the inset of Figure 3a, which clearly identifies (110), (101), (211), (220), and (301) diffraction rings of rutile SnO2.61 Detailed morphological information on the SnO2 nanowires is revealed by HRTEM characterization, as shown in Figure 3b. A lattice fringe spacing of ∼0.34 nm is figured out, which is consistent with the (110) plane of rutile SnO2. Figure 3c presents the XRD pattern of the SnO2 nanowires with wellresolved (110), (101), (111), (211), (220), (002), (221), and (112) crystallization peaks of rutile SnO2 (JCPDS card No. 411445). The (110) crystallization peak residing at 2θ = 26.7° is ascribed to a lattice fringe spacing of 0.33 nm, which is in good agreement with HRTEM observation (Figure 3b). As proven by BET measurement, the silicon hollow nanospheres have large surface areas (237.5 m2 g−1) which translate into extremely high surface energies and strong van der Waals forces. The organically modified SnO2 nanowires, once being introduced, will be spontaneously attracted by the
Figure 3. (a) TEM and (b) HRTEM images of SnO2 nanowires and (c) XRD pattern of SnO2 nanowires. The inset in (a) shows the ED pattern of SnO2 nanowires.
strong van der Waals forces and therefore facilely assemble on the naked surfaces of the silicon hollow nanospheres, forming micelle-like, stable Si@SnO2 core−shell heterostructures. In this way, the total free energy of the whole solution system is minimized by the formation of the SnO2 coating (shell). Figure 4a,b compares typical SEM images of the silicon hollow nanospheres before and after the self-assembly of the SnO2 nanowires. It can be seen that the naked surfaces of the silicon hollow nanospheres are uniformly decorated with the SnO2 nanowires after the self-assembly process. Therefore, the resulting Si@SnO 2 (weight ratio = 80:20) core−shell heterostructures, different from relatively smooth silicon hollow nanospheres, have a unique surface structure featured by a rough, porous SnO2 coating (shell). Note that there are very few SnO2 nanowires dissociated from the silicon hollow nanospheres, confirming a fairly high self-assembly efficiency. The overall morphology of the silicon hollow nanospheres is well retained, which demonstrates that our self-assembly approach based on van der Waals interactions is noninvasive imposing few, if any, defects to both building blocks. The SnO2 nanowires encapsulating the silicon hollow nanospheres form a conductive coating (shell) on the insulative silicon host material (core), which plays an important role in improving the reversibility of lithium storage.46 The SEM image of the Si@ SnO2 core−shell heterostructures is in good agreement with TEM characterization, and a typical TEM image is presented in Figure 4c. An HRTEM image of the Si@SnO2 core−shell heterostructures is given in Figure 4d, which clearly reveals the (110) plane of rutile SnO2 as well as the (111) plane of polycrystalline silicon. The coexistence of the two building blocks, namely, silicon and SnO2, is further vividly revealed by EDS mapping (Figure S3). A uniform distribution of elemental Si and Sn is observed corresponding to the profile of the Si@ SnO2 core−shell heterostructures. In the XRD pattern (Figure 4e), the crystallographic phases of coexisting silicon and SnO2 can be well indexed, with circles indicating (111), (220), (311), D
DOI: 10.1021/acsami.6b00107 ACS Appl. Mater. Interfaces XXXX, XXX, XXX−XXX
Research Article
ACS Applied Materials & Interfaces
Figure 4. (a) SEM image of silicon hollow nanospheres, (b−d) SEM, TEM, and HRTEM images of Si@SnO2 (weight ratio = 80:20) core−shell heterostructures, and (e) XRD pattern of Si@SnO2 (weight ratio = 80:20) core−shell heterostructures.
Figure 5. High-resolution Si 2p and Sn 3d XPS spectra of Si@SnO2 core−shell heterostructures.
the SnO2 nanowires are excess relative to the silicon hollow nanospheres. The chemical nature of the Si@SnO2 (weight ratio = 80:20) core−shell heterostructures is revealed by XPS characterization, as shown in Figure 5. As seen from the high-resolution Si 2p spectrum, the major peak centered at 99.8 eV is assigned to Si0, while the other deconvolved peaks located at higher binding energies are attributed to SiOx (0 < x ≤ 2).62 The presence of a certain degree of surface oxidation, as detected by XPS, is characteristic of silicon obtained by magnesiothermic reduction.63,64 However, it should be noted that the peak intensity of Si0 is much higher than that of silicon oxides, suggesting that the degree of surface oxidation is relatively low. Considering the fact that both TEM and XRD results rule out the presence of silicon oxides (Figure 2b,e), we can conclude that only the surfaces of the silicon hollow nanospheres are coated with a thin (
