DOI: 10.1002/ente.201300017 Si-Composite Anode for Lithium-Ion Batteries with High Initial Coulombic Efficiency Xingkang Huang, [a] Haejune Kim, [a] Shumao Cui, [a] Patrick T. Hurley, [b] and Junhong Chen* [a] Silicon as an anode material for lithium-ion batteries possess- es a high capacity of 4200 mAh g 1 , which attracts significant interest from the energy-storage research community; how- ever, its large volume change upon lithium insertion and ex- traction results in the pulverization of Si particles, very poor cycle performance, and low Coulombic efficiency, which hinder the practical application of Si in lithium-ion batter- ies. [1] To address these issues, great efforts have been made using various approaches. Using a constant-charge-capacity technique may help improve the cycle performance of Si anodes, [2–4] but this technique is inconvenient for practical ap- plications. Electrolyte additives such as vinylene carbonate, [5] fluoroethylene carbonate, [6] succinic anhydride, [7] and tris(- pentafluorophenyl) borane [8] were employed to improve the cycle performance at the cost of reducing the initial Coulom- bic efficiency. Si/C-composite anodes were proposed exten- sively with the assumption that the C in the composites serves as a buffer for the volume change of Si during charg- ing/discharging. [9–17] Employing nanowires, nanoparticles, thin films, and other nanostructures of Si is a common method to improve the cycle performance. [18–22] Forming Si alloys such as Si–Fe, [23–25] Si–Co, [26] Si–Ni, [27–29] and Si–Al–M (M = Cu, Fe, Mn, Ni, Cr), [30–33] is another approach used to enhance the cycle performance, but this also reduces the capacity of Si anodes. In this case, the content of active Si in the Si alloy was intentionally reduced, and MSi x phases (M = Fe, [25] Ni, [28] Co, [26] Mo, [34] etc.) served as inactive matrices to accommo- date the instantaneous volume change and to help improve the cycle performance upon lithium insertion and extraction. Here we report a Si–Fe–Al composite anode material, readi- ly synthesized by high-energy ball milling, with high initial Coulombic efficiency and excellent cycle performance. The content of Si in the alloy is strategically determined through experiments to achieve a balance between a high specific ca- pacity with significantly improved initial Coulombic efficien- cy and cycle performance for Si-based anodes. The as-prepared samples consisted of elemental Si uni- formly distributed in a matrix of inactive Si-based intermetal- lic phases. Figure 1 shows the XRD patterns of a mixture of Si, Fe, and Al powders and the as-prepared samples. Com- pared with the mixture of starting materials, sample SFA- 1 obtained by 10-hour ball milling shows that the Si-peak in- tensity was reduced significantly and some new broad peaks emerged (Figure 1a). The new peaks could be indexed to in- termetallic phases of FeSi 2 and Al 4.5 FeSi. The existence of the two intermetallic phases could be confirmed by compar- ing the XRD patterns of the three as-prepared samples as shown in Figure 1 b–d. As the weight percentage of Al in- creased, the content of Al 4.5 FeSi in the product increased and that of FeSi 2 decreased as indicated by the relative intensities of XRD peaks at approximately 47.88 and 48.88 (2 q). Mean- while, the peak at 28.48 still exists but with reduced peak in- tensity after ball milling, which suggests that there is still some elemental Si remaining in the as-prepared samples. Figure 2 shows a SEM image of a 12 mm particle from sample SFA-1 (Figure 2a) and its energy dispersive X-ray spectroscopy (EDS) elemental mapping (Figure 2 b–d), which suggests Si, Fe, and Al are well distributed in the as- prepared material. However, Si did not form solid-solution alloys with Al and Fe; instead, it formed intermetallic phases with Al and Fe due to the significant electronegativity differ- ence between Si and Al/Fe. Considering that the content of Si is more than necessary to completely form intermetallic phases, the remaining elemental Si is expected to distribute uniformly within the intermetallic matrix as illustrated in Fig- ure 3 a and confirmed by the high-resolution TEM [a] Dr. X. K. Huang, H. Kim, S. M. Cui, Prof. J. H. Chen Department of Mechanical Engineering University of Wisconsin-Milwaukee 3200 N Cramer Street, Milwaukee, WI 53211 (USA) E-mail: jhchen@uwm.edu [b] P. T. Hurley Global Technology & Innovation, Power Solutions Johnson Controls 5757 N Green Bay Avenue, Milwaukee, WI 53209 (USA) Figure 1. XRD patterns of (a) a mixture of Si, Fe, and Al powders, (b) the as- prepared sample SFA-1 (Fe 0.31 Al 0.14 Si 0.55 ) using the raw materials shown in (a), (c) sample SFA-2 (Fe 0.30 Al 0.15 Si 0.55 ), and (d) sample SFA-3 (Fe 0.29 Al 0.16 Si 0.55 ). The JCPDS numbers of patterns referred to Si, Fe, Al, FeSi 2 , and Al 4.5 FeSi are 27-1402, 6-696, 85-1327, 73-1843, and 82-546. Energy Technol. 2013, 1, 305 – 308  2013 Wiley-VCH Verlag GmbH& Co. KGaA, Weinheim 305