The requirements of the portable electronics and electric vehicle market have created a demand for lithium-ion battery (LIB) materials with high energy density. Silicon is a promising anode material yielding a high theoretical gravimetric capacity of 4200 mAh g-1.1 Earlier Si anode studies identified challenges introduced by Si’s large volume expansion during Li-Si alloying which leads to structure pulverization, electrode/binder contact failure and an unstable solid electrolyte interface, all of which result in reduced capacity and cyclability.1 These problems associated with volume expansion can be potentially addressed by nanostructuring (e.g., Si nano-particles or nanowires).1 While nanostructures can better accommodate the strain associated with volume expansion, they typically reduce tap density, require costly fabrication processes and can result in electrode degradation due to fragile nanostructures with subsequent cycling.2 Another approach that has been explored to suppress structural fracturing during cycling of Si anodes whilst maintaining their high capacity is to dope the Si.3–5 It has been proposed that volume expansion in intrinsic Si occurs because of poor electrical conductor and consequently lithiation only occurs at the surface leading to a large volume expansion of the surface region (leading to cracking and capacity loss) rather than extending into the fresh silicon.6 Dopants can be introduced into Si through chemical-vapour deposition processes or via the use of doped crystalline Si wafers. In many cases, the latter is the more convenient. McSweeney et al. use of lightly-B-doped (5-10 Ω cm) Si anodes results in a higher capacity and a crack free surface whereas, with more heavily-As-doped (0.001-0.005 Ω cm) Si, the capacity was reduced and the surface underwent extensive phase changes resulting in cracking.5 In both cases, lithiation-induced crystalline-to-amorphous phase changes occur at the wafer surface.7 However, with this study it was difficult to identify the individual contribution of dopant density and polarity in the observed differences in de-lithiation. The objective of this study is to investigate the effect of dopant density on Si anode capacity and capacity retention. N-type (100) double-sided polished Si wafer fragments (thickness 500 μm) with different dopant densities (10 Ω.cm (As), 0.01-0.05 Ω.cm (Sb) and 0.001-0.005 Ω.cm (As)) were assembled into coin cells with Li foils and 80 mL 1 M LiPF6 in EC/DMC 1:1 (v/v). Comparative nanostructured electrodes were also fabricated using metal assisted chemical etching (MACE) to create a nanowire electrode surface on one-side of the wafer. All electrodes were pre-lithiated at a current density of at 0.05 mA/cm² for either 10, 20 or 40 hrs by cycling between 2.5 V to 0.01 V (versus Li/Li+) prior to capacity measurements. Figure 1 A, B and C shows the capacity of the polished and MACE-treated 0.01-0.05 Ω.cm (Sb-doped) Si electrodes at a range of current densities after different initial pre-lithiation durations. Reversible capacity increased with pre-lithiation duration for all electrodes. The coulombic efficiency remained high for all the cycling rates tested indicating stable cycling. Although the MACE-treated Si electrodes resulted in higher areal capacities at low current density, for all pre-lithiation times their capacity reduced more with increased current density compared to the polished electrodes. This suggested that the nanowires formed on Si using MACE may have been damaged either through the long pre-lithiation phases or whilst subsequently being cycled. Figure 1 D shows the areal capacity as a function of cycle number for polished Si electrodes of different resistivity after 10 hrs pre-lithiation. All electrodes exhibited an initial increase in capacity ~ 50−100 cycles before capacity began to fade. Of interest is the observed capacity increase after ~ 250 cycles for the more heavily-doped electrodes. This may be due to increased Li alloying into the bulk of the Si wafers due to their higher conductivity, whereas for the least conductive sample, surface pulverization may have occurred due to the greater surface lithiation. These initial results suggest a complex interaction between Li and doped Si that may depend not only on the polarity of the dopants, but also their atomic properties and density in Si. The full paper will present extended cycling results and material investigations involving cycled electrodes. [1] Xu et al Progress in Materials Science 90, 1–44, 2017. [2] Shi et al. Nat. Commun. 7, 2016. [3] Domi et al. ACS Appl. Mater. Interfaces 8, 7125–7132, 2016. [4] Long et al. J. Phys. Chem. C 115, 18916–18921, 2011. [5] McSweeney et al. Electrochim. Acta 135, 356–367, 2014. [6] Szczech et al. Energy and Environmental Science 4, 56–72, 2011. [7] Limthongkul et al. Acta Mater. 51, 1103–1113, 2003. Figure 1
Read full abstract
Dec 1, 2019
Adnan Raza Khan + 2
Open Access