Synthesis of curved Si flakes using Mg powder as both the template and reductant and their derivatives for lithium-ion batteries

Abstract

Silicon has been considered as one of the most promising carbon-alternative anode materials for high-energy lithium-ion batteries (LIBs). However, it is still challenging to facilely prepare Si with unique micro-/nano-structures on a large scale. Here, we report a designed procedure for the synthesis of curved Si microflakes. Commercial Mg powder was used as both the active template for SiO₂ coating and reductant to subsequently reduce SiO₂ sheaths into Si by magnesiothermic reaction, for the first time. The as-obtained curved Si flakes could be further tailored. We demonstrated that Sn@carbon nanorods (CRNs) could be formed on the surface of curved Si flakes forming a Si@Sn@CNR composite. When tested as anode materials for LIBs, the as-prepared Si@Sn@CNR composites exhibited highly improved electrochemical performance as compared to pure Si materials.

---

1. Introduction

Lithium-ion batteries (LIBs) have been attracting much attention for decades. LIBs are considered the powerhouse of the popular mobile personal electronic devices, including cell phones and laptop computers.^1–3 LIBs are playing increasingly important roles in hybrid and electric vehicles.^4–6 To accommodate the growing demand for energy storage, tremendous efforts have been made to develop next-generation LIBs with improved electrochemical performances in terms of energy, power, cost and safety. The electrochemical performances of LIBs are determined by the electrode materials selected. Regarding the anode, carbonaceous materials are still dominantly used in commercial LIBs currently. However, graphite has a low theoretical capacity of 372 mA h g⁻¹ (by intercalation forming LiC₆). Therefore, next-generation LIBs should employ carbon-alternative negative electrode materials.

Among many potential carbon alternatives investigated, silicon is one of the most outstanding candidates. Silicon has the highest theoretical capacity of ∼3572 mA h g⁻¹ by forming Li_3.75Si based on alloying mechanism.^4,7–9 However, the insertion and extraction of lithium in and out silicon could lead to a very large volume variation (>300%), which cause cracking of materials and disintegration of the electrodes leading to capacity fading or poor cycling performance. To address the issue of volume variation, extensive efforts have been made to prepare nanoscale silicon. Various nanoscale silicon materials have been reported achieving certain level of success.^10,11 For example, 1-dimentional Si nanowires were grown on stainless steel substrate by vapor–liquid–solid growth method;⁹ 2-D Si nanofilms with thickness around 250 nm on Cu substrate prepared by magnetron sputtering were synthesized;¹² 3-D lotus-root-like Si was prepared through template assisted magnesiothermic reduction method.¹³ Silicon materials with unique morphologies have been prepared through various methods.¹² Si has been deposited on carbon nanowire through chemical vapor deposition method using SiH₄ to prepare carbon@silicon nanowire in argon at 500 °C.¹⁴ Template-assisted magnesiothermic reduction has also been applied to obtain Si from SiO₂. SBA-15 silica was used as silicon source and template to prepare 3-D mesoporous silicon.¹³ TiO₂ nanorods have been used as the template coated with SiO₂ and core–shell Ti@Si coaxial nanorods have been prepared by magnesiothermic reduction process.¹⁵ Graphene oxides have been applied as the substrate to grow SiO₂ particles, and 3-D porous structure of silicon/graphene has been prepared by magnesiothermic reduction process.¹⁶ However, it is still very challenging to achieve facile synthesis of Si materials in micro- or nano-scale at low cost by simple procedures.

Magnesiothermic reduction of SiO₂ to silicon is a promising method to effectively produce silicon with unique micro- or nano-structures.^5,13,16–19 For example, magnesiothermic reduction of micro-assemblies silica to obtain microporous nanocrystalline silicon replicas has been reported.¹⁷ 3-D porous Si was prepared through the magnesiothermic reduction of SiO₂ mesoporous powder.¹⁹ SiO₂ with unique structures as precursors normally have to be prepared by multi-step procedures, typically by template methods using metal, metal oxides or carbon based materials as sacrificed templates.²⁰ However, templates used for the preparation of silica precursors must be removed to get the silica precursors and the template removing process is typically destructive toward the products. We called those sacrificed templates as passive templates. Additional issue is that the sacrificed templates are difficult to be prepared. The complicated procedure with multiple steps and the use of expensive sacrificed templates make it challenging to mass produce silicon with unique structures. Therefore, it will be interesting to use active templates which can function as reductants at the same time.

Here, we developed a facile synthesis strategy where commercial magnesium powder was employed as multifunctional template for the coating of silica sheaths on it and also the reducing agent for magnesiothermic reduction of silica into curved silicon microflakes. We called the Mg templates as active templates. Other passive templates were not required. In other words, our synthesis strategy eliminated the complex template preparation and passive template-removing processes. Additionally, we also tried to improve the conductivity of the as-prepared silicon microflakes. Hierarchical structure of Si@Sn@CNRs to further improve the electrochemical performances of silicon microflakes was demonstrated. The preliminary results show that the electrochemical performance of Si@Sn@CNRs, which has a reversible capacity of 451 mA h g⁻¹ after 20 cycles, was greatly improved as compared to pristine Si microflakes, whose capacity dropped to 10 mA h g⁻¹ at the same testing condition. Fig. 1 shows the schematic of the design and procedure in the preparation of curved Si microflakes and Si@Sn@CNR composites.

Fig. 1 Schematic of the design and procedure in the preparation of curved Si flakes by using Mg powder as bifunctional template and reductant: (a) commercial Mg powder, (b) SiO₂ wrapped Mg template, (c) Si wrapped MgO obtained by magnesiothermic reduction of SiO₂ using Mg as the reductant, (d) curved Si flakes obtained by etching off MgO. Additional process to further tailor the curved Si flakes (e) SnO₂ nanoparticles decorated Si flake and (f) Si@Sn@CNRs high-order structures obtained by CVD treatment of (e).

2. Experimental section

All the reagents used were as received. The overall idea and experimental procedure are illustrated in Fig. 1.

Preparation of Mg@SiO₂

In a typical synthesis, 0.4 g of Mg powder (325 mesh) was added into 60 ml of 2-propanol under stirring, followed by 0.8 g of hexadecyltrimethyl-ammonium bromide (CTAB) and the mixture was stirred for 6 h. Then the Mg powder capped by CTAB were collected and re-dispersed into the mixture of 75 ml of 2-propanol, 16 ml of DI water and 4 ml of ammonia under stirring to form solution A. After 5 min of stirring, the mixture of 1 ml of tetraethyl orthosilicate (TEOS) and 45 ml of 2-propanol was added into solution A drop-wisely to form solution B. Solution B was kept under stirring for 12 hours. The precipitates were washed with water and ethanol, and then dried in a vacuum oven at room temperature for 24 h to obtain SiO₂ wrapped Mg microparticles or Mg@SiO₂. An excessive amount of Mg was purposely used here taking into consideration of possible certain fraction of Mg oxidized in the process to ensure enough Mg for reduction of SiO₂.

Preparation of Si microflakes

Mg@SiO₂ powder was put into a tube furnace, heated to 650 °C with ramping rate of 10 °C min⁻¹ in argon, and kept at 650 °C for another 4 hours for magnesiothermic reduction to get MgO@Si. The MgO@Si powder was washed by 2 M HCl solution and then HF acid solution (5 wt%) to remove MgO and SiO₂ remnant to obtain Si phase. The acid washed sample was then washed with water until the pH value was equal to the pH value of DI-water, and then washed with ethanol. The solid sample was dried at 80 °C in a vacuum overnight.

Preparation of pure Si@Sn@CNRs

Calculated amount of silicon microflakes (0.2384 g of Si (solid), 8.5 mmol) were put into 30 ml freshly prepared aqueous solution of SnCl₄ (0.1 ml of SnCl₄ (liquid), 0.85 mmol). The mixture was stirred for 24 h to allow the SnCl₄ to be hydrolyzed forming SnO₂ nanoparticles deposited on the Si microflakes. The Si microflakes decorated with SnO₂ were collected by centrifugation and dried at 80 °C in a vacuum oven. To grow CNRs on Si structures, the dried silicon microflakes decorated with SnO₂ nanoparticles were put into a crucible and inserted into a tube furnace. Argon was used to purge the tube furnace for 30 min to remove the air inside the tube furnace. Tube furnace was heated to 650 °C with ramping rate of 10 °C min⁻¹ under argon. The gas was switched to acetylene balanced with argon (10% of acetylene in the mixture) at rate of 100 sccm and maintained at 650 °C for 3 h for chemical vapor deposition (CVD) reactions. The furnace was then switched back to pure argon flow and cooled down in argon.

Materials characterization

X-ray diffraction (XRD) was carried out with Rigaku Smartlab X-ray diffractometer using Cu K_α radiation (λ = 0.15418 nm). The morphologies were characterized by scanning electron microscopy (JSM-6510LV SEM), equipped with X-ray energy-dispersive spectrometer (EDS), field emission scanning electron microscopy (JSM-7600 FESEM), and transmission electron microscopy (JEOL 2010 TEM, with accelerating voltage of 200 kV).

Electrochemical measurements

Typically, Si microflakes were mixed with conductivity enhancer (Super P carbon black, Timcal), polyvinylidene fluoride (PVDF) binder at 8 : 1 : 1 weight ratio in N-methylpyrrolidone (NMP) to form a homogenous slurry. The slurry of Si@Sn@CNRs microflakes was prepared similarly by mixing 90 wt% of the as prepared active materials and 10 wt% of polyvinylidene fluoride binder in N-methylpyrrolidone. Note that no conductivity enhancer of Super P carbon black was added. The slurry was then applied to copper discs as current collectors and dried in a vacuum oven at 100 °C for 24 h. Typically, the mass loading on the current collector for Si@Sn@CNRs is 8 mg per cm². CR2032-type coin cells were assembled in an argon-filled glove box using the coated copper disc as the working electrode, metallic lithium foil as the counter and reference electrodes, 1 M solution of LiPF₆ in a 50 : 50 v/v mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) as the electrolyte, and PP/PE/PP trilayer membranes (Celgard 2320) as the separator. The coin cells were galvanostatically charged and discharged in the voltage window of 0.01–2.5 V on a MTI BST8-WA battery tester at room temperature.

3. Results and discussion

The chemical changes and compositions of the intermediate samples outlined in Fig. 1 were characterized by XRD (Fig. 2). First, a layer of SiO₂ was coated on Mg template via the Stöber sol–gel process to obtain Mg@SiO₂ as the precursor. Tetraethyl orthosilicate (TEOS) in our experiment was hydrolyzed and SiO₂ formed in the presence of ammonia and water (Fig. 1b). This Stöber sol–gel process is popularly used for SiO₂ coating, such as SiO₂ coated metals (Au, Ag, Fe, Ni and so forth),²¹ SiO₂ coated metal oxides (iron oxide,²² cobalt oxide,²³ zinc oxide²⁴ and titanium oxide²⁵) and SiO₂ coated carbon.^20,26,27 To the best of our knowledge, this was the first time to coat SiO₂ on Mg microparticles as the active template. Fig. 2a shows the XRD pattern of silica wrapped magnesium powder (Mg@SiO₂). Only XRD peaks associated with metallic magnesium are observed (JCPDS Card no. 35-0821). No peaks associated with SiO₂ sheaths suggest the SiO₂ sheath is amorphous. SiO₂ sheath was subsequently reduced to Si through the magnesiothermic reduction process utilizing the Mg core as the reducing agent. Fig. 2b shows the XRD pattern of MgO@Si. All the peaks can be assigned to silicon (JCPDS Card no. 27-1402) and MgO (JCPDS Card no. 45-0946), indicating the successful reduction of amorphous SiO₂ into Si and the oxidization of Mg core to MgO. The MgO core could be easily etched off by dilute HCl. Additionally, dilute HF was used to etch off a trace amount of remnant of SiO₂ that might not fully reduced by magnesiothermic reduction. Fig. 2c shows the Si flakes after acid etching to remove MgO. All the peaks can be assigned to silicon (JCPDS Card no. 27-1402) and no peak for MgO is observed (Fig. 2c), indicating MgO was completely removed by etching. The Si flakes were further processed. First, SnO₂ nanoparticles were decorated on Si flakes through the hydrolysis of SnCl₄ in water to obtain Si@SnO₂ nanoparticles. Subsequently, the SnO₂ deposited on the Si flakes acted as catalysts for growth of CNRs via CVD. At the same time, the SnO₂ was reduced to metallic Sn under this CVD condition and Si@Sn@CNRs was obtained. The XRD pattern of the as-prepared flakes of Si@Sn@CNRs is shown in Fig. 2d. The peaks for Si can still be observed, and all other peaks can be assigned to metallic tin (JCPDS Card no. 86-2264). This indicates that the deposited SnO₂ were completely reduced to Sn during the CVD process.

Fig. 2 XRD patterns of (a) silica wrapped magnesium powder (Mg@SiO₂), (b) silicon wrapped magnesium oxide (MgO@Si), (c) silicon microflakes, and (d) silicon microflakes decorated with tin and carbon nanorods (Si@Sn@CNRs).

The morphologies of products obtained at different stage of process (Fig. 1a–d) were characterized by electron microscopes (Fig. 3). Fig. 3a and b shows the SEM image of commercial Mg powder at different magnification. The size of Mg powder is about 40 to 90 μm. The Mg powder is in the form of plates or bulk rods. The low and high magnification SEM images of SiO₂ wrapped Mg are shown in Fig. 3c and d, respectively. The overall size and morphology of the Mg@SiO₂ are almost the same as the Mg powder without SiO₂ wrapping. Fig. 3e, the low magnification SEM image shows the morphology of the MgO@Si obtained after the magnesiothermic reduction process. Compared to the Mg template or Mg@SiO₂, the size of the structure of the MgO@Si is reduced. The Mg was converted to MgO during the reaction and the volume of the core would change due to the different density of Mg (1.74 g cm⁻³ at room temperature) and MgO (3.58 g cm⁻³). Therefore, the Si shells could collapse into small pieces due to shrinking core. On the other hand, Mg template (also the reductant) may be melted due to its low melting point of 650 °C, left over the Si shells. The magnesiothermic reaction in reduction of SiO₂ into Si could involve both liquid and vapor phase Mg.²⁸ Fig. 3f shows the high magnification SEM image of the MgO@Si, where curved flake-like structure in micro scale can be observed. The Si microflakes obtained after acid etching to remove MgO and trace SiO₂ are shown in Fig. 3g and h. Curved Si flakes are observed reflecting the heritage of template used (Fig. 3h). These results evidenced that Mg powder play dual roles as both template and reductant in the formation of Si microflakes. We believe the same idea could be extended to achieve unique structured Si sheaths using Mg in other unique structures (e.g., Mg nanospheres, nanoplates, nanorods, nanourchins, etc.^29,30) as the active template and reductant to form desired Si structures.

Fig. 3 Scanning electron microscope (SEM) images at different magnification of (a and b) commercial Mg powder, (c and d) SiO₂ wrapped Mg powder, (e and f) Si wrapped MgO after magnesiothermic reduction process, and (g and h) curved Si flakes after acid etching.

To confirm the bifunctional roles played by the Mg powder as template and reductant, EDS analysis and elemental mapping were carried out for the intermediate compound of MgO@Si (Fig. 4). The composition of the MgO@Si after the magnesiothermic reduction process was Mg, O and Si, as revealed by EDS (Fig. 4a). The elements could be associated with the presence of MgO, Si and a trace amount of amorphous SiO₂, A section of Si sheath with MgO was further analyzed by elemental mapping (Fig. 4b). The mouth-like opening suggests the Mg core as template was melt and the Si sheath wrapping the core was broken after magnesiothermic reaction. The corresponding elemental mapping of the MgO@Si flake clearly shows the uniform element distribution of O (red), Mg (green) and Si (blue) in the structure. The broken of the freshly formed Si sheath could also be attributed to the MgO nanoparticles formed inside the Si sheath and MgO nanoparticles induced low mechanical stability.

Fig. 4 (a) EDS, (b) SEM image and corresponding elemental mapping of MgO@Si to show the even distribution of O (Red), Mg (Green) and Si (Blue).

The Si flakes were further tailored to make Si@Sn@CNR composites. Hybrid structures of silicon/carbon have been explored to improve the cycling performance of silicon anode. For example, silicon coated carbon nanofibers,¹⁴ carbon nanotubes decorated with Si nanoparticles,³¹ silicon@carbon core–shell nanowires,³² the silicon@carbon yolk–shell structures have been reported achieving certain level of success.³³ Here, we explored to make composite based on the curved Si flakes by covering the Si flakes with Sn@CNRs. There are four reasons to introduce Sn@CNRs: (1) carbon nanorods and carbon wrapping could dramatically improve the poor conductivity of Si flakes; (2) metallic Sn is not only highly conductive but also electrochemically active; (3) the nanorods could enhance the integration of the electrode by forming 1-D networks; (4) the voids between the 1-D nanorods can also accommodate the volume variation and alleviate the stress upon cycling. The SnO₂ nanoparticles decorated Si flakes are shown in Fig. 5a. The decorated SnO₂ nanoparticles acted as catalyst for the growth of carbon nanorods through chemical vapor deposition (CVD) process, which is well documented.^34,35 At the same time the SnO₂ nanoparticles were reduced to metallic Sn and wrapped inside the carbon nanorods, forming Sn@CNRs. The low magnification FESEM image of Si@Sn@CNRs shows that a large number of carbon nanorods formed on the surface of Si flakes (Fig. 5b). More details about the Sn@CNRs are revealed by magnified FESEM image in Fig. 5c. One can see that Sn@CNRs densely covered the Si flakes. The Sn@CNRs are about 200 nm in length with diameter around ∼70 nm. The detail of the Sn@CNRs is further revealed by TEM (Fig. 5d). The different contrast in the nanorods are associated with the light contrast carbon nanorods encapsulating dark contrast metallic Sn nanorods.^34–36 The EDS analysis of the as-prepared Si@Sn@CNRs confirmed its chemical composition (Fig. S1 in ESI†). The weight ratio of Si : Sn : C is 19 : 59 : 19. The sample will need to be further optimized to achieve better weight balanced between those elements and electrochemical performances.

Fig. 5 (a) FESEM image of SnO₂ nanoparticle decorated Si flakes, Si@SnO₂; FESEM images at (b) low magnification overall view and (c) high magnification view revealing the densely formed Si@Sn@CNRs; (d) TEM image of Si@Sn@CNRs flakes clearly showing the Sn@CNTs on the surface.

Electrochemical performances of the as-prepared Si flakes and the further tailored Si flakes with Sn@CNRs or Si@Sn@CNRs were thoroughly evaluated (Fig. 6). The charge–discharge profiles of both the as prepared curved Si flakes and the derived Si@Sn@CNRs were plotted in Fig. 6a and b, respectively. For Si flakes, a very long flat plateau around 0.1 V is observed in the first discharge or lithiation process (black line). This plateau contributes a capacity of ∼2250 mA h g⁻¹ which could be attributed to the formation of Li_xSi alloy.^5,10,37 The very high first cycle irreversible capacity loss of around 1000 mA h g⁻¹ has been widely observed for Si anode materials, which could be attributed the formation of solid electrolyte interface (SEI) layer and the irreversible Li⁺ trapping.^5,13,38,39 A slope ranging from 0.25 to 0.5 V was observed during the charge or de-lithiation process in the first two cycles, which could be associated with the de-alloy of Li_xSi.^5,10,40

Fig. 6 First two cycle charge–discharge profiles for (a) the as-prepared Si flakes and (b) the further tailored Si flakes of Si@Sn@CNRs, (c) dQ/dV for the first two cycles of Si@Sn@CNRs and (d) cyclic performance at 100 mA g⁻¹ for Si flakes (hollow points) and Si@Sn@CNRs (solid points).

In contrast to that of the as-prepared Si flake based anodes, no long plateau was observed in the charge–discharge profile of Si@Sn@CNRs (Fig. 6b). Instead, the capacity dropped much slower to around 0.5 V and a slope ranging from 0.5 to 0.01 V was observed in the lithiation process for the first two cycles. The improvement of the capacity retention demonstrated that the design of unique integrated structure of Si@Sn@CNRs and the decoration of conductive 1D Sn@CNR network can effectively enhance the electrochemical performance of Si flakes. The improved performances could be attributed to the decoration of Sn@CNRs on the surface of the Si flakes where the Sn@CNRs can enhance electric conductivity, alleviate the stress, and form a network to integrate the electrodes. The initial capacity of Si@Sn@CNRs of 950 mA h g⁻¹ is lower than that of pristine Si flakes, which is due to the lower theoretical capacity of both Sn and carbon as compared to the high theoretical capacity of Si of 3572 mA h g⁻¹. The theoretical specific capacity is estimated to be 1358 mA h g⁻¹ based on the weight percent of C, Si, and Sn in the composite based on EDS results. Due to the aforementioned advantages offered by the metallic tin and carbon, the fast capacity fading associated with silicon is not observed. The initial capacity loss decreased from 1000 mA h g⁻¹ for Si flakes to 200 mA h g⁻¹ for Si@Sn@CNRs flakes, which could be attributed to the dramatically reduced SEI formation. It is known that SEI formation on the surface of carbon is different from that of Si. The formation of stable beneficial SEI layer on carbon is well documented. The SEI on carbon can facilitate the transfer of Li⁺, prevent the formation of additional SEI in subsequent cycles and enhance cyclability.¹¹ In contrast, the SEI layer formed on Si is relatively unstable and can be cracked due to the huge volume variation upon cycling. The SEI can continuously form upon cycling with the exposed Si to electrolytes due to cracks leading to capacity fading.^33,41,42 Additionally, the Li⁺ trapping caused by poor conductivity of Si was significantly reduced with the presence of Sn@CNR network. The charge–discharge profiles for the first two cycles almost overlapped, indicating the same electrochemical reactions involved in the first two cycles, or dramatically improved cycling performance as compared to that of pristine Si flakes. To further understand the electrochemical reactions involved for the lithium storage in Si@Sn@CNRs, the differential capacity vs. voltage plots for the first two cycles were analyzed (Fig. 6c). The cathodic peaks at 0.6 V and 0.38 V and the anodic peaks between 0.6–0.9 V can be attributed to the lithium insertion into and extraction from Sn, respectively.⁴³ The cathodic peaks below 0.12 V and anodic peak at ∼0.48 V can attributed to the lithiation and de-lithiation process of silicon.^44,45 Meanwhile, the lithium insertion into carbon composition also contribute to part of the cathodic peak below 0.12 V.^46,47

The cycling performance of both Si flakes and Si@Sn@CNRs are compared in Fig. 6d. The pure Si (hollow symbols) shows the typical cycling performance of Si material with dramatically capacity fading upon cycling. The capacity fading of Si is well documented. For example, in the case of Si microparticles, the capacity fades from ∼3750 mA h g⁻¹ to ∼300 mA h g⁻¹;³⁸ And in another case of carbon-coated bulk Si, the capacity fades from 2630 mA h g⁻¹ to ∼300 mA h g⁻¹ after 30 cycles.^38,39 In our case of Si flakes, it has similar large capacity for the first cycle but the irreversible capacity loss is huge. The charge capacity faded from 791 mA h g⁻¹ in the 2^nd cycle to 138 mA h g⁻¹ after 5 cycles. As comparison, the compound of Si@Sn@CNRs has a smaller initial capacity, but the issue of fast capacity fading is significantly reduced. After 5 cycles, the specific charge capacity of Si@Sn@CNRs just slightly faded from 654 mA h g⁻¹ in the 2^nd cycle to 562 mA h g⁻¹, which is much better than that of pure Si. After 20 cycles, the specific charge capacity of Si@Sn@CNRs remains at 451 mA h g⁻¹. The charge capacity retention is 69.0%. In contrast the capacity of pure Si microflakes was declined to 10 mA h g⁻¹ after 20 cycles, with much smaller charge capacity retention at only 0.8%. The volumetric capacity and full-cell stack energy density were estimated to be about 647 A h l⁻¹ and 710 W h l⁻¹.⁴⁸ The values estimated are slightly lower but still comparable to that of graphite. However, our materials have not been optimized and only 30% of the theoretical capacity was released in our preliminary study. We believe that further effort could optimize and dramatically improve their performances. The coulombic efficiency of Si@Sn@CNRs was also shown in Fig. 6d, it kept increasing for the first 10 cycles and remained at above 98% from the 10^th cycle, which may due to the stabilization of SEI layer and the activation process of active materials.⁴⁹ Practically, the coulombic efficiency obtained for our Si@Sn@CNRs sample is still not as good as required and further optimization is still needed. Addition to optimization of electrode materials, there are a few options in the preparation of the electrode and cell testing processes, including annealing PVDF binder instead of vacuum drying alone, using carboxymethyl cellulose or lithium polyacrylate as binders instead of PVDF, putting additives (such as fluoroethylene carbonate) in the conventional electrolytes, electrode activation, which are under our investigation. Here, we tried to compare the two samples tested under the same conditions. As comparison, the increasing and stabilization of coulombic efficiency of pure Si flakes are slower, which may due to the pulverization of Si and the continuous formation of SEI layer in the first few cycles. As compared to the pristine Si flakes, the cycling performance of Si@Sn@CNRs microflakes is significantly improved. Both samples were further tested over 60 cycles at various currents (Fig. S2 in the ESI†). The bare Si flakes quickly faded to insignificant capacity upon further cycling at increased currents. In contrast, the Si@Sn@CNRs microflakes could still achieve a reasonable capacity without dramatic capacity fading when the testing current was increased. The improved performances successfully verified our hypothesis discussed previously in the adoption of Sn@CNRs in to Si flakes to enhance their electrochemical performances.

4. Conclusion

In summary, a procedure for synthesis of curved Si flakes was developed. Commercial micro-size magnesium powder (∼325 mesh) was used as bifunctional template to support the coating of SiO₂ sheath and as reductant to subsequently reduce the SiO₂ into Si via magnesiothermic reaction. Additionally, we also demonstrate a facile post-synthesis procedure to further tailor the as-prepared Si flakes by coating with Sn@CNRs making an unique composite of Si@Sn@CNRs high-order structures. Electrochemical evaluation suggests that the not-optimized Si@Sn@CNRs composite demonstrated improved Li ion storage performances. Our ongoing effort is to further investigate the effect of different Mg active template with uniquely defined structures, including various nanostructures to synthesize unique Si nanostructures, and further optimize Si@Sn@CNR composites. We believe this approach could be used to synthesize Si with different nanostructures. However, the high cost will be another limitation of our approach which is worthy for further investigation and improvement.

Acknowledgements

The authors thank Guangzhao Mao for XRD instrument and the Lumigen Instrument Center for SEM and TEM instruments, Wayne State University, Detroit, MI.

References

1. J. M. Tarascon and M. Armand, Nature, 2001, 414, 359–367.
2. M. Armand and J. M. Tarascon, Nature, 2008, 451, 652–657.
3. B. Dunn, H. Kamath and J.-M. Tarascon, Science, 2011, 334, 928–935.
4. P. G. Bruce, B. Scrosati and J.-M. Tarascon, Angew. Chem., Int. Ed., 2008, 47, 2930–2946.
5. F. Cheng, J. Liang, Z. Tao and J. Chen, Adv. Mater., 2011, 23, 1695–1715.
6. L. Ji, Z. Lin, M. Alcoutlabi and X. Zhang, Energy Environ. Sci., 2011, 4, 2682–2699.
7. F. Liu, S. Song, D. Xue and H. Zhang, Nanoscale Res. Lett., 2012, 7, 1–17.
8. A. S. Arico, P. Bruce, B. Scrosati, J.-M. Tarascon and W. van Schalkwijk, Nat. Mater., 2005, 4, 366–377.
9. C. K. Chan, H. Peng, G. Liu, K. McIlwrath, X. F. Zhang, R. A. Huggins and Y. Cui, Nat. Nanotechnol., 2008, 3, 31–35.
10. U. Kasavajjula, C. Wang and A. J. Appleby, J. Power Sources, 2007, 163, 1003–1039.
11. J. R. Szczech and S. Jin, Energy Environ. Sci., 2011, 4, 56–72.
12. J. P. Maranchi, A. F. Hepp and P. N. Kumta, Electrochem. Solid-State Lett., 2003, 6, A198–A201.
13. H. Jia, P. Gao, J. Yang, J. Wang, Y. Nuli and Z. Yang, Adv. Energy Mater., 2011, 1, 1036–1039.
14. L.-F. Cui, Y. Yang, C.-M. Hsu and Y. Cui, Nano Lett., 2009, 9, 3370–3374.
15. X. Meng and D. Deng, ACS Appl. Mater. Interfaces, 2015, 7, 6867–6874.
16. X. Xin, X. Zhou, F. Wang, X. Yao, X. Xu, Y. Zhu and Z. Liu, J. Mater. Chem., 2012, 22, 7724–7730.
17. Z. Bao, M. R. Weatherspoon, S. Shian, Y. Cai, P. D. Graham, S. M. Allan, G. Ahmad, M. B. Dickerson, B. C. Church, Z. Kang, H. W. Abernathy Iii, C. J. Summers, M. Liu and K. H. Sandhage, Nature, 2007, 446, 172–175.
18. K. H. Sandhage and Z. Bao, U.S. Pat., 7615206 B2, 2009.
19. Y. Yu, L. Gu, C. Zhu, S. Tsukimoto, P. A. van Aken and J. Maier, Adv. Mater., 2010, 22, 2247–2250.
20. J. Liu and X.-W. Liu, Adv. Mater., 2012, 24, 4097–4111.
21. S. Liu and M. Y. Han, Chem.–Asian J., 2010, 5, 36–45.
22. S. Kralj, D. Makovec, S. Čampelj and M. Drofenik, J. Magn. Magn. Mater., 2010, 322, 1847–1853.
23. Y. Meng, D. Chen and X. Jiao, J. Phys. Chem. B, 2006, 110, 15212–15217.
24. Y. Chen, X. Xue and T. Wang, Nanotechnology, 2005, 16, 1978.
25. Y. Ren, M. Chen, Y. Zhang and L. Wu, Langmuir, 2010, 26, 11391–11396.
26. K. Ding, B. Hu, Y. Xie, G. An, R. Tao, H. Zhang and Z. Liu, J. Mater. Chem., 2009, 19, 3725–3731.
27. Y. Zhu, Y. Pan, H. Xu, C. Xiang, H. Kou and J. Guo, Chem. Lett., 2007, 36, 1098–1099.
28. T.-D. Nguyen, J. A. Kelly, W. Y. Hamad and M. J. MacLachlan, Adv. Funct. Mater., 2015, 25, 2175–2181.
29. W. Li, C. Li, C. Zhou, H. Ma and J. Chen, Angew. Chem., Int. Ed., 2006, 45, 6009–6012.
30. B. Peng, J. Liang, Z. Tao and J. Chen, J. Mater. Chem., 2009, 19, 2877–2883.
31. A. Gohier, B. Laïk, K.-H. Kim, J.-L. Maurice, J.-P. Pereira-Ramos, C. S. Cojocaru and P. T. Van, Adv. Mater., 2012, 24, 2592–2597.
32. H. Kim and J. Cho, Nano Lett., 2008, 8, 3688–3691.
33. N. Liu, H. Wu, M. T. McDowell, Y. Yao, C. Wang and Y. Cui, Nano Lett., 2012, 12, 3315–3321.
34. D. Deng, M. G. Kim, J. Y. Lee and J. Cho, Energy Environ. Sci., 2009, 2, 818–837.
35. D. Deng and J. Y. Lee, J. Mater. Chem., 2010, 20, 8045–8049.
36. Y. Zou and Y. Wang, ACS Nano, 2011, 5, 8108–8114.
37. L.-F. Cui, L. Hu, J. W. Choi and Y. Cui, ACS Nano, 2010, 4, 3671–3678.
38. C. Wang, H. Wu, Z. Chen, M. T. McDowell, Y. Cui and Z. Bao, Nat. Chem., 2013, 5, 1042–1048.
39. H. Yoo, J.-I. Lee, H. Kim, J.-P. Lee, J. Cho and S. Park, Nano Lett., 2011, 11, 4324–4328.
40. A. Rong, X. P. Gao, G. R. Li, T. Y. Yan, H. Y. Zhu, J. Q. Qu and D. Y. Song, J. Phys. Chem. B, 2006, 110, 14754–14760.
41. B. Liu, P. Soares, C. Checkles, Y. Zhao and G. Yu, Nano Lett., 2013, 13, 3414–3419.
42. K. Fu, O. Yildiz, H. Bhanushali, Y. Wang, K. Stano, L. Xue, X. Zhang and P. D. Bradford, Adv. Mater., 2013, 25, 5109–5114.
43. W.-M. Zhang, J.-S. Hu, Y.-G. Guo, S.-F. Zheng, L.-S. Zhong, W.-G. Song and L.-J. Wan, Adv. Mater., 2008, 20, 1160–1165.
44. M. K. Datta and P. N. Kumta, J. Power Sources, 2009, 194, 1043–1052.
45. A. M. Chockla, J. T. Harris, V. A. Akhavan, T. D. Bogart, V. C. Holmberg, C. Steinhagen, C. B. Mullins, K. J. Stevenson and B. A. Korgel, J. Am. Chem. Soc., 2011, 133, 20914–20921.
46. J. Y. Eom, H. S. Kwon, J. Liu and O. Zhou, Carbon, 2004, 42, 2589–2596.
47. G. T. Wu, C. S. Wang, X. B. Zhang, H. S. Yang, Z. F. Qi, P. M. He and W. Z. Li, J. Electrochem. Soc., 1999, 146, 1696–1701.
48. M. N. Obrovac and V. L. Chevrier, Chem. Rev., 2014, 114, 11444–11502.
49. J. Hassoun, F. Bonaccorso, M. Agostini, M. Angelucci, M. G. Betti, R. Cingolani, M. Gemmi, C. Mariani, S. Panero, V. Pellegrini and B. Scrosati, Nano Lett., 2014, 14, 4901–4906.

---

Footnotes

† Electronic supplementary information (ESI) available: EDS and charge–discharge cycling test over 60 cycles under various currents. See DOI: 10.1039/c5ra10218a

‡ D.D. designed the research. J.Z. conducted the experiments. D.D. and J.Z. analyzed the results and wrote the manuscript.

---