From Photovoltaic Si Sludge to High‑Performance Li‑Ion Battery Anodes via Oxygen Diffusion

Background

Lithium‑ion batteries are central to modern portable electronics and electric vehicles (EVs) [1]. The growing demand for high‑energy, low‑cost, and safe batteries has intensified research into alternative anode materials. Silicon, with its highest theoretical capacity among known elements, attracts considerable attention. Unfortunately, lithium insertion into silicon triggers a volumetric expansion exceeding 300 %, leading to particle pulverization, loss of electrical contact, and rapid capacity fade [3]. Numerous strategies—silicon nanowires, porous silicon, multi‑shell composites, and binder‑free designs—have been explored to mitigate these issues [4–8]. Yet, most of these approaches depend on expensive, low‑yield commercial silicon nanoparticles (e.g., Sigma‑Aldrich product 795585) [10].

Photovoltaic manufacturing generates substantial amounts of silicon sludge—crystalline fragments of 1–2 µm that possess a high surface area conducive to oxidation. Annually, the industry produces over 100,000 MT of such sludge, representing an untapped resource for energy storage applications. Prior attempts to repurpose sludge have involved complex processes, including nickel plating and graphene CVD growth [10]. Solid sub‑oxide silicon (e.g., SiO) has also shown promise as an anode, forming Li₂O and Li₄SiO₄ matrices that dampen silicon’s expansion [11,12]. These findings highlight the importance of controlled oxygen incorporation in tailoring silicon anode structures.

Methods

Silicon sludge was sourced from LONGI Silicon Materials Corp. and cleaned with HCl and ethane to remove surface contaminants. The sludge, which typically exhibits a flake morphology due to cleavage along the tetrahedral silicon lattice, was annealed in alumina crucibles under ambient air at 550 °C for 10 h. This step promotes oxygen inward diffusion, transforming the bulk silicon into a nano‑Si/SiO_x composite (brownish appearance, see Fig. 1c). The annealed powder (1 g) was dispersed in 240 mL deionized water, followed by the addition of 0.8 mL ammonium hydroxide (28 % NH₃·H₂O). After vigorous stirring for 20 min, 400 mg resorcinol and 0.56 mL 37 % formaldehyde were introduced and stirred overnight to form a resorcinol–formaldehyde (RF) resin coating. The RF layer was carbonized at 850 °C under argon (heating rate 5 °C min⁻¹, 2 h) to yield a carbon shell. Finally, HF (10 wt %) was used to etch the SiO_x, producing the Si/C yolk/shell structure. A control sample—identical processing without the initial oxygen diffusion step—was also prepared for comparison. The overall workflow is illustrated in Fig. 1a.

Electrochemical characterization employed 1 M LiPF₆ in a 1:1:1 (v/v/v) EC/DEC/DMC electrolyte with a Celgard 2400 separator. Working electrodes consisted of 80 wt % active material, 10 wt % acetylene black, and 10 wt % PVDF binder in NMP. Half‑cells were cycled between 0.01 V and 2.5 V at 100 mA g⁻¹ using a LAND CT2001A platform. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were performed on a CHI660C workstation at 0.5 mV s⁻¹, with an AC amplitude of 10 mV over 10⁵ – 0.01 Hz.

Results and Discussion

X‑ray photoelectron spectroscopy (XPS) revealed the oxidation states of silicon in the samples (Fig. 2). The pristine sludge displayed predominantly elemental Si (Si⁰), while the nano‑Si/SiO_x product exhibited significant sub‑oxide peaks (Si¹⁺–Si⁴⁺), confirming successful oxygen incorporation (Table 1). X‑ray fluorescence (XRF) quantified the weight fractions of Si and O, showing that extended annealing increased the SiO_x content (Table 2).

Transmission electron microscopy (TEM) and selected area electron diffraction (SAED) (Fig. 3a,b) confirmed the coexistence of crystalline silicon islands within an amorphous oxide matrix, characteristic of the nano‑Si/SiO_x structure. STEM/EDX mapping (Fig. 3c) further demonstrated the miscibility of Si and O at the nanoscale. After HF etching, a clear yolk/shell Si/C composite emerged (Fig. 3d). While HF is traditionally considered hazardous, industry practices for recycling fluorosilicic acid can render the process greener [19].

Electrochemical testing (Fig. 4) showcased the superiority of the nano‑Si/SiO_x‑derived Si/C electrode. Over 500 cycles at 100 mA g⁻¹, the capacity remained above 1,250 mAh g⁻¹, with an average Coulombic efficiency of 99.5 %. In contrast, the control sample failed after fewer than 20 cycles, underscoring the critical role of the SiO_x layer and subsequent yolk/shell architecture. Voltage profiles (Fig. 4b) displayed a typical SEI‑forming plateau near 0.75 V during the first cycle and a stable 0.5 V plateau associated with Li–Si de‑alloying in subsequent cycles. CV curves (Fig. 4c) confirmed persistent Li alloying/de‑alloying peaks, indicating durable silicon participation. EIS (Fig. 4d) revealed low charge‑transfer resistance, attributable to the conductive carbon shell.

Conclusions

We have established a scalable, cost‑effective route to convert silicon sludge—a by‑product of PV wafer fabrication—into a high‑performance LIB anode. By harnessing oxygen inward diffusion, followed by carbon coating and selective oxide removal, a Si/C yolk/shell composite was produced that delivers exceptional cycling stability and capacity. This “win‑win” approach not only valorizes silicon waste but also provides a low‑cost alternative to conventional nano‑silicon anodes, benefiting both the photovoltaic and energy‑storage sectors.