Carbon-Free Conversion of SiO2 to Si via Ultra-Rapid Alloy Formation: Toward the Sustainable Fabrication of Nanoporous Si for Lithium-Ion Batteries

Silicon has the potential to improve lithium-ion battery (LIB) performance substantially by replacing graphite as an anode. The sustainability of such a transformation, however, depends on the source of silicon and the nature of the manufacturing process. Today’s silicon industry still overwhelmingly depends on the energy-intensive, high-temperature carbothermal reduction of silica—a process that adversely impacts the environment. Rather than use conventional thermoreduction alone to break Si–O bonds, we report the efficient conversion of SiO2 directly to Mg2Si by a microwave-induced Mg plasma within 2.5 min at merely 200 W under vacuum. The underlying mechanism is proposed, wherein electrons with enhanced kinetics function readily as the reductant while the “bombardment” from Mg cations and electrons promotes the fast nucleation of Mg2Si. The 3D nanoporous (NP) Si is then fabricated by a facile thermal dealloying step. The resulting hierarchical NP Si anodes deliver stable, extended cycling with excellent rate capability in Li-ion half-cells, with capacities several times greater than graphite. The microwave-induced metal plasma (MIMP) concept can be applied just as efficiently to the synthesis of Mg2Si from Si, and the chemistry should be extendable to the reduction of multiple metal(loid) oxides via their respective Mg alloys.


Porosity (%) =
(S1) The BJH desorption cumulative volume of pores (between 1.7 -300.0 nm in diameter) was 0.262711 cm 3 g -1 . This volume was employed as the Cumulated Pore Volume in equation (S1). The volume of Si was calculated as 1 g of Si divided by its density. According to Table S3 and equation (S1), the calculated porosity is 37.97%.
Transmission X-ray Microscopy (TXM) characterization was conducted at the beamline station of Taiwan Light Source (TLS 01B1), National Synchrotron Radiation Center (NSRRC) in Hsinchu, Taiwan. [8] The Ge (111) toroidal focusing mirror provided monochromatic light with a photon energy of 8 keV. The transmitted beam passed through a zone-plate and a phase ring to generate an image. The phase ring was positioned at the back focal plane of the zone plate which recorded phase-contrast images at the charge-coupled device (CCD) detector. The beam size for sample observation is about 1 mm × 0.4 mm with an average photon flux of 3×10 11 photon．sec -1 ．200 mA -1 . The spatial resolution and field of view of TXM is 50 nm and 15 × 15 μm 2 , respectively. TXM 2D tomographic images were collected with a camera binning of 512 × 512 in pixels in the duration of 60 s exposure time. The collected TXM images were further processed and analyzed using ImageJ. The Faproma-alignment algorithm was adopted to correct the vertical and rotational motion errors along each projection to improve 3D reconstructed images. In addition, a maximum likelihood estimation reconstruction method was applied on the 3D image reconstruction using 151 sequential projections along a specific azimuth angle rotation (-90 − 90°). The visible 3D tomographic images and video were reconstructed using Amira 3D image processing software.

Estimation of the Energy Consumption of MIMP vs. Conventional Mg 2 Si Synthesis on a Lab Scale.
A rough estimate of the lab scale energy consumption can be performed for the MIMP reduction-magnesiation process of SiO 2 to Mg 2 Si, based on the prototype bench-top setup described in the Experimental section in the main paper and Figure S1. This therefore initially assumes a yield of 1 mmol Si per experiment. The equation of the MIMP process is given by: To Based on Equation (i), no CO x gases are chemically emitted during the MIMP process. The conventional high temperature synthesis process from SiO 2 to Mg 2 Si, as summarized in the main paper proceeds via: Here, Reaction (ii) is a high temperature carbothermal reduction (⩾ 1900 °C) and Reaction (iii) is a high-temperature solid-state synthesis under flowing inert gas (e.g. in a tube furnace).
To synthesize 5 g of Mg 2 Si, Reaction (ii) would typically utilize a 6 kW electrical induction or arc furnace for 1 h thus consuming 6 kW h of electricity and also emitting ~ 130 mmol of CO (~ 3.7 g). Reaction (iii) could be performed with a tube furnace (3.6 kW) under flowing Ar gas for 10 h, thus consuming 36 kW. The entire process (ii + iii) hence requires approximately 42 kW h of energy (and emits CO gas). (Using commercial Si in reaction (iii) would reduce the energy consumption based on an energy cost of 12 kW h kg -1 to produce Si from SiO 2 industrially, as described in Ref 33 in the main paper).
Although the current bench-top setup is only at the prototype stage, the results demonstrate the promise of scaling up the MIMP process. Further scale-up can potentially be achievable through alternative commercial and/or bespoke reactors with larger and/or multiple cavities and optimized designs, on the foundation of which, the energy efficiency of the MIMP physiochemistry will promisingly further increase. Figure S1. Experimental setup for the MIMP synthesis.      Figure S7. The visualization of frames in the crystal structure of SiO 2 (quartz) (model adopted from [3]). Atoms: red -oxygen, royal blue -Si.