Silicon is an important semiconductor, used in many electronic components. In addition, it also finds application in conversion of solar energy and electrochemical energy storage. The high theoretical capacity of more than 4 Ah/g makes Si an attractive alloying material for anodes in Lithium-Ion Batteries (LIB) [1]. However, one major drawback of the Si-based materials is the complicated and expensive fabrication process. Among the other methods, electrodeposition is a low-cost and simple alternative because it allows adjusting the properties of the silicon layer by varying the electrochemical parameters. Elemental silicon is highly reactive, for which reason the deposit can react with the substrate and/or the electrolyte. The aim of this work is to investigate the substrate influence on the electrochemistry of the silicon deposition from an Ionic Liquid (IL) and to analyze its impact on the composition of the resulting layers. As highly stable and low volatile solvents, ionic liquids offer the opportunity for an effective silicon electrodeposition. Cu and Ni were selected as substrates with the perspective to further apply them as porous hosts for Si in LIBs. For comparison, vitreous carbon (GC) was used as an inert material. The electrolyte consisted of 0.5 M SiCl4 dissolved in 1-butyl-1-methyl-pyrrolidinium bis(trifluoromethylsulfonyl)imide [BMP][TFSI]. The electrochemical reduction of SiCl4 was studied with linear sweep voltammetry (LSV). Chronoamperometry, coupled with electrochemical quartz crystal microbalance (QCM) was used to investigate the mass-charge balance of the deposition process. Scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy (XPS) were applied for the analysis of the surface and the composition of the obtained layers. The voltammetric data exhibited just one peak for vitreous carbon, indicating a one-step reduction of SiCl4 (fig. 1). In contrast, three peaks for Cu and two peaks and two shoulders for nickel were observed. Simultaneously performed QCM measurements did not show any frequency change above -1.3 V in the case of nickel, which indicates that the first reduction processes (E > -1.3V) are not correlated with any bulk deposition, but could be related to a partial reduction of the precursor to soluble species SixCl4-y. A similar behavior can be observed for the copper substrate. Until -1.0 V the frequency of the quartz crystal did not change and the following decrease is so small, that bulk deposition can be ruled out. The partial reduction of SiCl4 to a soluble species SixCly and/or adsorption phenomena on the electrode surface can be an explanation for this observation. At more cathodic potentials (E < -1.3 V) the resonance frequency of the QCM decreases, which corresponds to a mass increase at the electrode surface. The evaluation of the potentiostatic EQCM data results in MQCM/Cu = 65.2 g/mol and MQCM/Ni = 52.5 g/mol, assuming a transfer of four electrons. This is quite different compared to the theoretical value of Mtheo = 28.1 g/mol and can be related to an entrapment of the electrolyte or its decomposition products. However, after 60 minutes the frequency change reached a steady state for both substrates accompanied with a strong damping increase. The deposition is inhibited at this point, which is probably due to the low conductivity of the deposited layer. The XPS measurements confirm the assumption about the entrapment of IL into the deposits. Furthermore, the data exhibit a high reactivity of silicon resulting in the formation of metal silicides at the interface for Cu and Ni. This supposition is underpinned by the absence of the signal for Si0 for samples, which were stored under argon for several weeks before the XPS measurements (fig. 1). Nevertheless, this can enhance the adhesion of the deposit. The performances of these silicon layers in LIBs were investigated by galvanostatic cycling in EC/DMC (1:1) with 10% FEC and 1 M LiPF6. [1] Vlaic, Codruta Aurelia, et al. "Electrochemical lithiation of thin silicon based layers potentiostatically deposited from ionic liquid." Electrochimica Acta 168 (2015): 403-413. Figure 1
Read full abstract