Porous Si/Fe2O3 Dual Network Anode for Lithium–Ion Battery Application

Benefiting from ultra-high theoretical capacity, silicon (Si) is popular for use in energy storage fields as a Li–ion battery anode material because of its high-performance. However, a serious volume variation happens towards Si anodes in the lithiation/delithiation process, triggering the pulverization of Si and a fast decay in its capacity, which greatly limits its commercial application. In our study, a porous Si/Fe2O3 dual network anode was fabricated using the melt-spinning, ball-milling and dealloying method. The anode material shows good electrochemical performance, delivering a reversible capacity of 697.2 mAh g−1 at 200 mA g−1 after 100 cycles. The high Li storage property is ascribed to the rich mesoporous distribution of the dual network structure, which may adapt the volume variation of the material during the lithiation/delithiation process, shorten the Li–ion diffusion distance and improve the electron transport speed. This study offers a new idea for developing natural ferrosilicon ores into the porous Si-based materials and may prompt the development of natural ores in energy storage fields.


Introduction
With the fast development of the market economy, the requirements for energy storage devices are constantly improving [1][2][3]. Lithium-ion batteries (LIBs) are identified as attractive energy maintaining devices because of their high specific energy, long service life and excellent environmental compatibility [4][5][6]. At present, the low theoretical capacity of the graphite anode restricts its commercial application in the high-end market [7,8]. To enhance the properties of LIBs further, it is necessary to develop new anodes with high performance. Silicon (Si) is considered to be one of the potential candidates to substitute conventional carbon anode for next-generation LIBs on account of the characteristics of extremely high theoretical mass-specific capacity (4200 mAh g −1 ), low working voltage and high natural abundance [9,10]. The main obstacle to its practical application is the tremendous volume variation (≈420%) during repeated lithiation/delithiation procedures [11], bring about pulverization, shedding of active materials and rapid decay of capacity [12][13][14]. To accommodate the volume variation of Si and boost the electrochemical properties of LIBs, some strategies have been proposed by scholars from various countries. Firstly, once the Si materials were produced into different nanomaterials, such as method was used to analyze the specific surface and the Barrett-Joyner-Halenda (BJH) method was employed to determine the pore size distribution.
Si/Fe 2 O 3 material, carboxymethyl cellulose (CMC) and conductive agent (Super P) (8:1:1, mass ratio) were mixed in deionized water to create a slurry. The slurry was smeared onto the surface of the Cu current collector (Cu foil) and dehydrated in a vacuum furnace at 60 • C for 12 h and then cut into discs of 10 mm in diameter to make a working anode. The mass loading of the anode was in the range of 0.89-1.05 mg/cm 2 . In a glove box with an argon environment (H 2 O/O 2 , less than 0.01 PPM), the anode, the lithium tablet cathode, the Celgard 2320 diaphragm and the electrolyte (1 M LiPF 6 dissolved in EC/DEC, 1:1 in volume ratio) were encapsulated in a CR2032 coin battery case. The electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) curves of the cell were detected through an electrochemical workstation (CHI760E) with a test scan speed of 0.1 mV s −1 in 0.01-3 V. Galvanostatic charge/discharge curves were monitored on a newway battery tester in the range of 0.01-3 V (vs. Li + /Li). Nanomaterials 2020, 10, x FOR PEER REVIEW 3 of 12 Brunauer-Emmett-Teller (BET) method was used to analyze the specific surface and the Barrett-Joyner-Halenda (BJH) method was employed to determine the pore size distribution. Si/Fe2O3 material, carboxymethyl cellulose (CMC) and conductive agent (Super P) (8:1:1, mass ratio) were mixed in deionized water to create a slurry. The slurry was smeared onto the surface of the Cu current collector (Cu foil) and dehydrated in a vacuum furnace at 60 °C for 12 h and then cut into discs of 10 mm in diameter to make a working anode. The mass loading of the anode was in the range of 0.89-1.05 mg/cm 2 . In a glove box with an argon environment (H2O/O2, less than 0.01 PPM), the anode, the lithium tablet cathode, the Celgard 2320 diaphragm and the electrolyte (1 M LiPF6 dissolved in EC/DEC, 1:1 in volume ratio) were encapsulated in a CR2032 coin battery case. The electrochemical impedance spectroscopy (EIS) and cyclic voltammetry (CV) curves of the cell were detected through an electrochemical workstation (CHI760E) with a test scan speed of 0.1 mV s −1 in 0.01-3 V. Galvanostatic charge/discharge curves were monitored on a newway battery tester in the range of 0.01-3 V (vs Li + /Li).

Results and Discussion
XRD results of the as-obtained ball-milled samples and the dealloyed samples are presented in Figure 2a, [29][30][31]. With the extension of the ball-milling time, a part of the Al elements may dissolve into the FeSi phase, inducing a broader peak. In addition, the relative intensity of the FeSi diffraction peak increases gradually while that of the Al peak declines, indicating that the elemental distribution has greatly changed, as shown in Figure 2a. No diffraction peak of Al can be found in the dealloyed materials (Figure 2b), indicating that the most of Al elements have been leached out. In this situation, Fe elements are oxidized concurrently in the dealloying process by oxygen dissolving in corrosion liquid, generating Fe2O3. Besides the diffraction peak of Fe2O3, diffraction peaks of (111), (200) and (311) crystal planes of Si phases are also observed [32][33][34]. The above results confirm that the as-synthesized dealloying product is mainly composed of Si and Fe2O3 phases.

Results and Discussion
XRD results of the as-obtained ball-milled samples and the dealloyed samples are presented in Figure 2a,b, respectively. After ball-milling for a different time, the samples present strong diffraction peaks at 38 [29][30][31]. With the extension of the ball-milling time, a part of the Al elements may dissolve into the FeSi phase, inducing a broader peak. In addition, the relative intensity of the FeSi diffraction peak increases gradually while that of the Al peak declines, indicating that the elemental distribution has greatly changed, as shown in Figure 2a. No diffraction peak of Al can be found in the dealloyed materials (Figure 2b), indicating that the most of Al elements have been leached out. In this situation, Fe elements are oxidized concurrently in the dealloying process by oxygen dissolving in corrosion liquid, generating Fe 2 O 3 . Besides the diffraction peak of Fe 2 O 3 , diffraction peaks of (111), (200) and (311) crystal planes of Si phases are also observed [32][33][34]. The above results confirm that the as-synthesized dealloying product is mainly composed of Si and Fe 2 O 3 phases.  Figure 2c displays the Raman spectra of the dealloyed products, the peaks observed at 520 cm −1 correspond to Si-Si bonds [35], while the peaks at about 293 cm −1 relate to the Fe-O bonds (Fe2O3) [36]. Furthermore, the peaks in the range of 900-1000 cm −1 are in accord with the multi-phonon peaks of Si [37]. The Raman data indicate that the products mainly contain two components of Fe oxides and Si, which is compatible with the XRD result.  Figure 3a-c. BM-24-4 presents a typical network-like structure consisting of ligaments and pores. It can be observed that the surface of the ligaments is loaded with nanosheets. A large number of nanosheets with a length of ~140 nm are connected together, but they block the pores of the network, which is not conducive to the passage of lithium ions during charging and discharging process. Figure 3b presents the dual network structure of BM-48-4, which is composed of the nanoparticles (diameter: ≤50 nm) accumulation network and the nanosheets (length of ~500 nm) network. The nanoparticles fill the interspace among the nanosheets network, improving the utilization rate of space. The two sets of networks interpenetrate each other to form a double network structure, which is beneficial to buffer the volume variation in the repeated charging/discharging procedure. Figure 3c shows the microstructure of the BM-72-4 sample. It can be seen that the product is composed of a blocky structure (accumulated by coarsened nanoparticles) and the thick nanosheets network, showing a composite with poor porous structure.  Figure 2c displays the Raman spectra of the dealloyed products, the peaks observed at 520 cm −1 correspond to Si-Si bonds [35], while the peaks at about 293 cm −1 relate to the Fe-O bonds (Fe 2 O 3 ) [36]. Furthermore, the peaks in the range of 900-1000 cm −1 are in accord with the multi-phonon peaks of Si [37]. The Raman data indicate that the products mainly contain two components of Fe oxides and Si, which is compatible with the XRD result.  Figure 3a-c. BM-24-4 presents a typical network-like structure consisting of ligaments and pores. It can be observed that the surface of the ligaments is loaded with nanosheets. A large number of nanosheets with a length of~140 nm are connected together, but they block the pores of the network, which is not conducive to the passage of lithium ions during charging and discharging process. Figure 3b presents the dual network structure of BM-48-4, which is composed of the nanoparticles (diameter: ≤50 nm) accumulation network and the nanosheets (length of~500 nm) network. The nanoparticles fill the interspace among the nanosheets network, improving the utilization rate of space. The two sets of networks interpenetrate each other to form a double network structure, which is beneficial to buffer the volume variation in the repeated charging/discharging procedure. Figure 3c shows the microstructure of the BM-72-4 sample. It can be seen that the product is composed of a blocky structure (accumulated by coarsened nanoparticles) and the thick nanosheets network, showing a composite with poor porous structure.
The microstructure of the optimal BM-48-4 sample was further analyzed by TEM. The interlacing distribution of nanoparticles and finely fragmented nanosheets demonstrates that the nanoporous structure is formed across the composite as shown in Figure 3d The formation process of iron oxide can be explained as follows. In the NaOH solution, OH − groups violently collide with the Al-Fe-based ribbons to form the intermediate Fe(OH)2 [38]. Then, this phase decomposes into Fe, Fe3O4 and H2. At the same time, the reduction of Fe3O4 by H2 happens, inducing the formation of Fe2O3 and even α-Fe in the dealloying products. The composition of the dealloyed product is affected by the concentration of the corrosive solution, the proportion of the initial element, the ratio between the reactant and the corrosive solution, etc. In a more concentrated NaOH solution (e.g., 10 M), the reaction between Al and NaOH is enhanced, producing H2 in a short period. Most of the Fe3O4 is reduced and thus α-Fe will dominate in the reaction products. While in NaOH solutions with appropriate concentrations (e.g., 1-2 M), iron oxides will dominate in the reaction products.
Nitrogen adsorption-desorption isotherm and pore diameter distribution curves are displayed in Figure 4a,b, respectively. The results show that the specific surface area of BM-48-4 (38.4 m 2 g −1 ) is higher than that of BM-24-4 (23.9 m 2 g −1 ) and BM-72-4 (11.2 m 2 g −1 ). A large number of mesopores less than 10 nm exist in the product (Figure 4b), which can cushion the volume expansion of the The formation process of iron oxide can be explained as follows. In the NaOH solution, OH − groups violently collide with the Al-Fe-based ribbons to form the intermediate Fe(OH) 2 [38]. Then, this phase decomposes into Fe, Fe 3 O 4 and H 2 . At the same time, the reduction of Fe 3 O 4 by H 2 happens, inducing the formation of Fe 2 O 3 and even α-Fe in the dealloying products. The composition of the dealloyed product is affected by the concentration of the corrosive solution, the proportion of the initial element, the ratio between the reactant and the corrosive solution, etc. In a more concentrated NaOH solution (e.g., 10 M), the reaction between Al and NaOH is enhanced, producing H 2 in a short period. Most of the Fe 3 O 4 is reduced and thus α-Fe will dominate in the reaction products. While in NaOH solutions with appropriate concentrations (e.g., 1-2 M), iron oxides will dominate in the reaction products.
Nitrogen adsorption-desorption isotherm and pore diameter distribution curves are displayed in Figure 4a,b, respectively. The results show that the specific surface area of BM-48-4 (38.4 m 2 g −1 ) is higher than that of BM-24-4 (23.9 m 2 g −1 ) and BM-72-4 (11.2 m 2 g −1 ). A large number of mesopores less than 10 nm exist in the product (Figure 4b), which can cushion the volume expansion of the material and shorten the Li-ion diffusion distance. Based on the above analysis, BM-48-4 is expected to possess relatively good electrochemical performance [39]. The surface elements and valence states of BM-48-4 were analyzed by XPS. The full XPS spectrum shown in Figure 4c shows that elements Si, Fe, Al, O and C exist in BM-48-4 without other impure substances. The appearance of Al 2p spectra stems from un-dealloyed residual Al elements. The Si 2p spectrum in Figure 4d reveals two typical peaks of Si located at 98.3 and 102.1 eV, relating to Si 0 and Si 4+ , respectively. This indicates that slight oxidation occurs in the outermost layer of Si [40]. Fe 2p spectrum in Figure 4e shows clear peaks concentrated at 710.8, 724.8 and 719.8 eV, which are connected to Fe 2p 3/2 , Fe 2p 1/2 and satellite peaks, respectively. In addition, an energy difference of 14.0 eV between Fe 2p 3/2 and Fe 2p 1/2 can be obtained, demonstrating the generation of Fe 2 O 3 [41,42]. The O 1s spectrum shown in Figure 4f can be divided into two peaks, the peak at 531.8 eV is from hydroxyl (OH bond, possibly from residual sodium hydroxide), and the wide peak centered at 530.8 eV can be attributed to the peak of metal bond in oxide, namely Fe-O bond (OM bond) in Fe 2 O 3 [43].
of Al 2p spectra stems from un-dealloyed residual Al elements. The Si 2p spectrum in Figure 4d reveals two typical peaks of Si located at 98.3 and 102.1 eV, relating to Si 0 and Si 4+ , respectively. This indicates that slight oxidation occurs in the outermost layer of Si [40]. Fe 2p spectrum in Figure 4e shows clear peaks concentrated at 710.8, 724.8 and 719.8 eV, which are connected to Fe 2p3/2, Fe 2p1/2 and satellite peaks, respectively. In addition, an energy difference of 14.0 eV between Fe 2p3/2 and Fe 2p1/2 can be obtained, demonstrating the generation of Fe2O3 [41,42]. The O 1s spectrum shown in Figure 4f can be divided into two peaks, the peak at 531.8 eV is from hydroxyl (OH bond, possibly from residual sodium hydroxide), and the wide peak centered at 530.8 eV can be attributed to the peak of metal bond in oxide, namely Fe-O bond (OM bond) in Fe2O3 [43]. To study the charging/discharging reaction mechanism of Si/Fe2O3 anode in detail, a CV test was performed towards the BM-48-4 electrode with the scanning speed of 0.1 mV s -1 in the range of 0.01~3.0 V, as presented in Figure 5a. In the initial reduction process, an apparent peak at 0.2 V and a steep peak appeared at 0.01~0.15 V can be found, relating to the creation of amorphous LixSi from lithiation of crystal Si and the generation of solid electrolyte interphase (SEI) film. Two obvious peaks of 0.3 V and 0.5 V correspond to delithiation reaction from LixSi to Si [44,45] in the first anode scan. The peak appearing at about 1.1 V corresponds to the reaction between the surface oxygenic functional groups and Li + [46]. A relatively wide peak at 1.85 V is in connection with the oxidation of metal Fe to Fe 2+ /Fe 3+ and decomposition of Li2O. From the second cycle, two peaks can be found at 1.3 V and 0.68 V in the cathode reaction, which are believed to the multi-step electrochemical reaction from Fe2O3 to Fe (Fe2O3→LixFe2O3→Li2Fe2O3→Fe) [47]. While the broad peak in the anode reaction is decomposed into two peaks at 1.65 V and 1.85 V, relating to the transformation from Fe 0 to Fe 2+ and from Fe 2+ to Fe 3+ , respectively. In addition, the CV curves in different cycles reflect relatively good reversibility [48]. As the number of cycles increases, the closure area of CV curves decreases slightly, showing that the BM-48-4 composite possesses acceptable cyclic stability. Although there have been a lot of studies on the application of Fe2O3 anode in LIBs, there is some controversy on its electrochemical reaction mechanism, and opinions have not been unified yet. A lot of detection and analysis is still needed in the future. To study the charging/discharging reaction mechanism of Si/Fe 2 O 3 anode in detail, a CV test was performed towards the BM-48-4 electrode with the scanning speed of 0.1 mV s −1 in the range of 0.01~3.0 V, as presented in Figure 5a. In the initial reduction process, an apparent peak at 0.2 V and a steep peak appeared at 0.01~0.15 V can be found, relating to the creation of amorphous Li x Si from lithiation of crystal Si and the generation of solid electrolyte interphase (SEI) film. Two obvious peaks of 0.3 V and 0.5 V correspond to delithiation reaction from Li x Si to Si [44,45] in the first anode scan. The peak appearing at about 1.1 V corresponds to the reaction between the surface oxygenic functional groups and Li + [46]. A relatively wide peak at 1.85 V is in connection with the oxidation of metal Fe to Fe 2+ /Fe 3+ and decomposition of Li 2 O. From the second cycle, two peaks can be found at 1.3 V and 0.68 V in the cathode reaction, which are believed to the multi-step electrochemical reaction from Fe 2 O 3 to Fe (Fe 2 O 3 →Li x Fe 2 O 3 →Li 2 Fe 2 O 3 →Fe) [47]. While the broad peak in the anode reaction is decomposed into two peaks at 1.65 V and 1.85 V, relating to the transformation from Fe 0 to Fe 2+ and from Fe 2+ to Fe 3+ , respectively. In addition, the CV curves in different cycles reflect relatively good reversibility [48]. As the number of cycles increases, the closure area of CV curves decreases slightly, showing that the BM-48-4 composite possesses acceptable cyclic stability. Although there have been a lot of studies on the application of Fe 2 O 3 anode in LIBs, there is some controversy on its electrochemical reaction mechanism, and opinions have not been unified yet. A lot of detection and analysis is still needed in the future.   29.5%, which is close to that of BM-24-4 (29.4%) and much lower than that of BM-72-4 (37.8%). With the increase in cycle number, the profile gradually moves to the left, demonstrating a slight capacity decline after several cycles. The charge/discharge profiles of BM-48-4 for the 50th and 100th cycles are close, revealing good cycling stability of the anode at the later stage of cycling. BM-24-4 and BM-72-4 electrodes show relatively low reversible capacity after cycling for 100 cycles. The huge capacity decay is induced by the serious volume variation during the cycling process, the crushing of active Si particles, the cracking of nanosheets structure, the creation of over-thick SEI film and the failure of electron and ion transport channels during the cycling process [49]. The difference in the electrochemical performance of a material in different tests in this work may be caused by the local non-uniformity of a material.
The rate performance of three electrode materials were tested at different current densities in the range of 200 to 5000 mA g −1 , as presented in Figure 6a. The BM-48-4 electrode delivers the reversible capacity of 1356.5, 963.1, 779.9, 604.3 and 512.2 mAh g −1 under 200, 500, 1000, 2000 and 5000 mA g −1 , respectively. At each current density, the BM-48-4 electrode shows the best specific capacity in the three anodes. When the current density recovers to 500 mA g −1 , the BM-48-4 anode delivers a reversible capacity of 906 mAh g −1 after 30 cycles, which is extremely higher than the other two anodes, revealing a relatively good rate property of BM-48-4 in three electrodes. Figure  6b presents the charge and discharge curves of the BM-48-4 electrode at different current densities.  Taking BM-48-4 as an example, the first discharge curve presents three platforms, which are 1.2~1.5, 0.6~1.0 and 0.01~0.2 V, respectively. While in the first charging stage, two long platforms of 0.2~0.7 and 1.1~2.0 V are found, according to the CV results. The capacity loss of the first cycle of BM-48-4 is about 29.5%, which is close to that of BM-24-4 (29.4%) and much lower than that of BM-72-4 (37.8%). With the increase in cycle number, the profile gradually moves to the left, demonstrating a slight capacity decline after several cycles. The charge/discharge profiles of BM-48-4 for the 50th and 100th cycles are close, revealing good cycling stability of the anode at the later stage of cycling. BM-24-4 and BM-72-4 electrodes show relatively low reversible capacity after cycling for 100 cycles. The huge capacity decay is induced by the serious volume variation during the cycling process, the crushing of active Si particles, the cracking of nanosheets structure, the creation of over-thick SEI film and the failure of electron and ion transport channels during the cycling process [49]. The difference in the electrochemical performance of a material in different tests in this work may be caused by the local non-uniformity of a material.
The rate performance of three electrode materials were tested at different current densities in the range of 200 to 5000 mA g −1 , as presented in Figure 6a. The BM-48-4 electrode delivers the reversible capacity of 1356.5, 963.1, 779.9, 604.3 and 512.2 mAh g −1 under 200, 500, 1000, 2000 and 5000 mA g −1 , respectively. At each current density, the BM-48-4 electrode shows the best specific capacity in the three anodes. When the current density recovers to 500 mA g −1 , the BM-48-4 anode delivers a reversible capacity of 906 mAh g −1 after 30 cycles, which is extremely higher than the other two anodes, revealing a relatively good rate property of BM-48-4 in three electrodes. Figure 6b presents the charge and discharge curves of the BM-48-4 electrode at different current densities. It shows that with the enhancement in current density, the curve gradually shifts to the left, that is, the capacity of the electrode gradually reduces. Moreover, when the current density recovers from high current density to 500 mA g −1 , the constant current charge/discharge profile almost coincides with the original curve tested at the same current density. All tests uncover that the BM-48-4 electrode possesses the best rate performance. Nanomaterials 2020, 10, x FOR PEER REVIEW 8 of 12 It shows that with the enhancement in current density, the curve gradually shifts to the left, that is, the capacity of the electrode gradually reduces. Moreover, when the current density recovers from high current density to 500 mA g −1 , the constant current charge/discharge profile almost coincides with the original curve tested at the same current density. All tests uncover that the BM-48-4 electrode possesses the best rate performance.   (Figure 7a) reduces, and the nanoparticles reunite to agglomeration, which is not conducive to ion transport. The BM-48-4 material (Figure 7b) can maintain the original structure of nanosheets and nanoparticles (with local agglomeration), presenting a relatively good structural stability. It is observed from Figure 7c that the surface of BM-72-4 is rough, and there are large cracks and some local aggregation. The above phenomenon reveals that BM-48-4 best preserves the original structure, which guarantees its good cyclic stability. Figure 7d,e shows the EIS data of Si/Fe2O3 anodes before and after 100 cycles. The EIS is composed of the high and medium frequency region (concave semicircle) caused by charge transfer resistance and the low-frequency region (slash) because of ion diffusion [50]. The minimum semicircle of BM-48-4 shows that the transfer resistance of the anode is significantly inferior to that of the BM-24-4 and BM-72-4 electrodes (Figure 7d). The inclination of the BM-48-4 electrode is higher than that of the BM-24-4 and BM-72-4 electrodes, indicating that the diffusion resistance of the BM-48-4 electrode is smaller. After 100 cycles, the diffusion resistance of three anodes is similar while the BM-48-4 electrode remains the smallest transfer resistance in three anodes (Figure 7e). Figure 7f shows a digital photo of a yellow light emitting diode (LED) bulb powered by an as-assembled half-cell. After 30 minutes, the LED bulb was less bright (Figure 7g) than it was when it started, but it still works, showing its good potential in practical applications. Table 1 [14,17,19,20,34,44,51] compares the electrochemical properties of the electrode materials currently studied with those previously reported. The Li storage performance of the as-synthesized Si/Fe2O3 anode is better than that of most listed Si-based composites, and its excellent electrochemical performance is mainly attributed to the following points. Firstly, Si/Fe2O3 electrodes with a high specific surface area may provide a large area of contact and interaction between the active material and the electrolyte. Secondly, the three-dimensional porous network structure with interconnected ligaments can enhance ionic mobility and permeability. Moreover, ample pores can effectively alleviate the volume expansion of Si. In summary, Si/Fe2O3 material synthesized from natural ferrosilicon ores in this study possesses immense potential as an anode for LIBs application. Furthermore, the study also offers a new idea for the synthesis of low-cost Si-based electrodes and opens a new direction for the application of ferrosilicon ores in a brand-new field other than deoxidizing agents for steel manufacturing.  After a long cycle, the pore size and porosity of BM-24-4 ( Figure 7a) reduces, and the nanoparticles reunite to agglomeration, which is not conducive to ion transport. The BM-48-4 material (Figure 7b) can maintain the original structure of nanosheets and nanoparticles (with local agglomeration), presenting a relatively good structural stability. It is observed from Figure 7c that the surface of BM-72-4 is rough, and there are large cracks and some local aggregation. The above phenomenon reveals that BM-48-4 best preserves the original structure, which guarantees its good cyclic stability. Figure 7d,e shows the EIS data of Si/Fe 2 O 3 anodes before and after 100 cycles. The EIS is composed of the high and medium frequency region (concave semicircle) caused by charge transfer resistance and the low-frequency region (slash) because of ion diffusion [50]. The minimum semicircle of BM-48-4 shows that the transfer resistance of the anode is significantly inferior to that of the BM-24-4 and BM-72-4 electrodes (Figure 7d). The inclination of the BM-48-4 electrode is higher than that of the BM-24-4 and BM-72-4 electrodes, indicating that the diffusion resistance of the BM-48-4 electrode is smaller. After 100 cycles, the diffusion resistance of three anodes is similar while the BM-48-4 electrode remains the smallest transfer resistance in three anodes (Figure 7e). Figure 7f shows a digital photo of a yellow light emitting diode (LED) bulb powered by an as-assembled half-cell. After 30 minutes, the LED bulb was less bright (Figure 7g) than it was when it started, but it still works, showing its good potential in practical applications. Table 1 [14,17,19,20,34,44,51] compares the electrochemical properties of the electrode materials currently studied with those previously reported. The Li storage performance of the as-synthesized Si/Fe 2 O 3 anode is better than that of most listed Si-based composites, and its excellent electrochemical performance is mainly attributed to the following points. Firstly, Si/Fe 2 O 3 electrodes with a high specific surface area may provide a large area of contact and interaction between the active material and the electrolyte. Secondly, the three-dimensional porous network structure with interconnected ligaments can enhance ionic mobility and permeability. Moreover, ample pores can effectively alleviate the volume expansion of Si. In summary, Si/Fe 2 O 3 material synthesized from natural ferrosilicon ores in this study possesses immense potential as an anode for LIBs application. Furthermore, the study also offers a new idea for the synthesis of low-cost Si-based electrodes and opens a new direction for the application of ferrosilicon ores in a brand-new field other than deoxidizing agents for steel manufacturing. Nanomaterials 2020, 10, x FOR PEER REVIEW 9 of 12

Conclusions
Porous Si/Fe2O3 dual network material was synthesized through the melt-spinning, ball-milling and dealloying process by utilizing natural ferrosilicon ores. The as-obtained Si/Fe2O3 anode displays a dual network structure. Due to this special porous structure, the BM-48-4 anode uncovers good Li storage property, delivering a good reversible capacity of 697.2 mAh g −1 after cycling at 200 mA g −1 for 100 cycles, which shows its great potential as LIBs anode. In addition, we also offer a new idea for the synthesis of porous Si-based electrodes and open a new direction for the application of ferrosilicon ores in a brand-new field.

Conclusions
Porous Si/Fe 2 O 3 dual network material was synthesized through the melt-spinning, ball-milling and dealloying process by utilizing natural ferrosilicon ores. The as-obtained Si/Fe 2 O 3 anode displays a dual network structure. Due to this special porous structure, the BM-48-4 anode uncovers good Li storage property, delivering a good reversible capacity of 697.2 mAh g −1 after cycling at 200 mA g −1 for 100 cycles, which shows its great potential as LIBs anode. In addition, we also offer a new idea for the synthesis of porous Si-based electrodes and open a new direction for the application of ferrosilicon ores in a brand-new field.

Funding:
The authors would like to acknowledge the financial support from the Science and Technology Project of Hebei Education Department, China (ZD2018059) and the Natural Science Foundation of Hebei Province, China (E2020202071, B2019202277).