In the pursuit of higher energy density for lithium-ion batteries, silicon-based materials have emerged as promising anode candidates due to their exceptional theoretical capacity. However, the severe volume expansion during lithiation poses significant challenges, including electrode pulverization and unstable solid-electrolyte interphase (SEI) formation. The formation process is a critical step that dictates the quality of the SEI layer, which in turn profoundly influences battery performance metrics such as Coulombic efficiency, rate capability, and cycle life. This study delves into the optimization of formation protocols for lithium-ion batteries incorporating silicon oxide (SiO)-graphite composite anodes, leveraging insights from SEI formation mechanisms, kinetic and thermodynamic analyses, and experimental validation.
The solid-electrolyte interphase is a passivating layer that forms on the anode surface during the initial charging cycles. Its composition and structure are determined by the reduction of electrolyte components. For graphite anodes, the SEI is typically composed of inorganic compounds like Li2CO3 and LiF closer to the electrode, and organic polymeric species outwardly. The formation process for a graphite-based lithium-ion battery is relatively well-understood. In contrast, the SEI on silicon-based materials is more complex. The native oxide layer (SiOx) on silicon particles reacts readily with electrolyte, forming organosilicon compounds. Upon reduction, this layer generates species that can further decompose solvents, leading to a dynamically evolving SEI. The large volume changes in silicon (up to 300-400%) during cycling cause continuous fracture and reformation of the SEI, consuming active lithium and electrolyte, which accelerates capacity fade. Therefore, engineering a robust, flexible, and stable SEI through a tailored formation process is paramount for the success of silicon-containing anodes in commercial lithium-ion battery applications.
The kinetics of lithium insertion differ markedly between graphite and silicon oxide. Graphite possesses high electronic conductivity (102–103 S cm-1) and moderate Li+ diffusion coefficients (10-6–10-11 cm2 s-1). Silicon oxide, however, has much lower electronic conductivity (10-5–10-4 S cm-1) and significantly slower ionic diffusion (10-14–10-13 cm2 s-1). This disparity can be expressed using Fick’s law for diffusion and Ohm’s law for electronic conduction. The overall current during formation is a combination of Faradaic and non-Faradaic processes. The Faradaic current related to SEI formation and lithium intercalation can be described by the Butler-Volmer equation:
$$ i = i_0 \left[ \exp\left(\frac{\alpha_a F \eta}{RT}\right) – \exp\left(-\frac{\alpha_c F \eta}{RT}\right) \right] $$
where \( i \) is the current density, \( i_0 \) is the exchange current density, \( \alpha \) are transfer coefficients, \( F \) is Faraday’s constant, \( \eta \) is overpotential, \( R \) is the gas constant, and \( T \) is temperature. The low exchange current density for SiO reactions implies that its lithiation is more polarization-sensitive. The formation overpotential \( \eta_{\text{form}} \) is crucial; high overpotentials (e.g., from high currents) can lead to excessive electrolyte reduction and a porous, inorganic-rich SEI, while moderate overpotentials favor a more organic, flexible layer.
Thermodynamically, the lithiation potentials of graphite and silicon oxide differ. Graphite intercalates lithium at potentials close to 0.1 V vs. Li/Li+, while silicon alloys with lithium at potentials below 0.5 V. In a full cell with a high-nickel NCM cathode, these correspond to specific voltage windows. To design an effective formation process, we performed Galvanostatic Intermittent Titration Technique (GITT) measurements on half-cells to map the voltage-capacity (V-Q) profiles. The results for the composite anode (87 wt% graphite, 13 wt% SiO) are summarized below. The voltage hysteresis between charge and discharge is indicative of polarization, which is more pronounced for the SiO component.
| Process | Voltage Range (vs. Li/Li+) | Approx. Full Cell Voltage | Capacity Contribution (Graphite : SiO) | Dominant Reaction |
|---|---|---|---|---|
| Lithiation Stage 1 | 0.7 V – 0.2 V | ~3.8 V – 4.0 V | 0.31 : 0.69 | Initial SEI formation, partial SiO lithiation |
| Lithiation Stage 2 | 0.2 V – 0.005 V | ~4.0 V – 4.2 V | 0.77 : 0.23 | Graphite intercalation, deep SiO alloying |
| Delithiation Stage 1 | 0.005 V – 0.2 V | ~4.2 V – 4.0 V (discharge) | 0.97 : 0.03 | Graphite de-intercalation dominates |
| Delithiation Stage 2 | 0.2 V – 1.5 V | ~4.0 V – 2.75 V (discharge) | 0.25 : 0.75 | SiO de-alloying dominates |
This thermodynamic map is vital. It shows that in the high-voltage range (4.0-4.2 V for the full cell), both materials are active during charge, but SiO contributes significantly only in the later stage of charge. During discharge, graphite de-lithiates first. A formation strategy that includes shallow cycling in this high-voltage window could promote simultaneous, controlled SEI formation on both materials while managing polarization.
Based on these principles, we designed and investigated four distinct formation protocols for 7300 mAh pouch-type lithium-ion batteries with NCM811 cathodes and SiO-graphite composite anodes. The cells were manufactured using standard electrode fabrication, stacking, vacuum drying, electrolyte filling, and sealing processes. The electrolyte was a standard LiPF6-based carbonate mixture with additives. After a 24-hour wetting stand, the cells underwent formation under controlled pressure. The four protocols are detailed below.
| Protocol ID | Core Strategy | Step-by-Step Charging Procedure (Temperature: 45°C unless noted) | Theoretical Rationale |
|---|---|---|---|
| FP-1 (Baseline) | Stepwise Constant Current (CC) | 1. 0.05C CC to 3.8 V (120 min) 2. 0.1C CC to 4.2 V (300 min) 3. 0.2C CC to 4.2 V (60 min) 4. Rest 2 min, then discharge at 0.2C to 3.0 V for aging. |
Gradual current increase allows sequential SEI layer growth: a dense inner layer followed by a more porous outer layer, balancing ionic conductivity and electronic insulation. |
| FP-2 (High-Voltage Shallow Cycle) | CC-CV with Shallow Cycling at High Voltage | 1. Rest 30 min. 2. 0.05C CC to 3.8 V (120 min). 3. 0.1C CC to 4.2 V (300 min). 4. 0.2C CC to 4.2 V, then CV at 4.2 V to 0.05C cutoff (full charge). 5. Rest 2 min. 6. 0.2C discharge to 3.9 V (~80% SOC). 7. Repeat steps 4-6 for a total of 3 shallow cycles (4.0V-4.2V range). 8. Final discharge to 3.0 V. |
Promotes deep lithiation of SiO in the high-voltage window, allowing volume expansion and SEI formation under controlled conditions. The cycling stabilizes the interface before full-depth cycling. |
| FP-3 (High Current) | Aggressive CC Steps | 1. 0.05C CC to 3.8 V (120 min) 2. 0.2C CC to 4.2 V (150 min) 3. 0.3C CC to 4.2 V (40 min) 4. Rest 2 min, then discharge at 0.2C to 3.0 V. |
Tests the impact of higher current density on SEI nucleation rate. Expected to form a thicker, possibly more inorganic and brittle SEI due to faster reduction kinetics. |
| FP-4 (Elevated Temperature) | Stepwise CC at Higher Temperature | Same step sequence as FP-1, but conducted at 60°C. | Higher temperature increases ion mobility and reaction rates, potentially leading to a more complete but possibly less stable SEI with higher organic solvent co-intercalation risk. |
All protocols were followed by standard aging, degassing, and capacity grading steps. The formed cells were then subjected to a comprehensive electrochemical characterization suite to evaluate the SEI quality and overall cell performance.
The voltage profile during the formation process itself provides initial insights. The cumulative charge capacity during the first charge was very similar for FP-1, FP-2, and FP-4, around 7800 mAh. FP-3 showed a slightly lower charge capacity (approx. 7700 mAh), suggesting that the higher currents may have led to slightly different reaction pathways or earlier termination due to voltage limits. The discharge capacity after formation, representing the reversible capacity, was highest for FP-2. The first-cycle Coulombic efficiency (CE) is a direct indicator of irreversible lithium loss, primarily to SEI formation. The results were: FP-2: 92.5%, FP-1: 91.8%, FP-4: 91.5%, FP-3: 90.9%. The superior first CE of FP-2 indicates that its formation protocol minimized parasitic reactions, likely by forming a more efficient and stable SEI layer in a controlled manner.
A key metric for storage performance is the voltage drop rate, or K-value. It measures the open-circuit voltage (OCV) decay over time and reflects the stability of the SEI against electron tunneling and micro-shorts. The K-value was calculated as:
$$ K = \frac{OCV(t_3) – OCV(t_2)}{t_3 – t_2} \quad \text{(in mV/h)} $$
where OCV measurements were taken at 30% SOC after 24 hours (t2) and 144 hours (t3). The results clearly ranked the protocols: FP-2 (0.0296 mV/h) < FP-1 (0.0353 mV/h) < FP-4 (0.0455 mV/h) < FP-3 (0.0568 mV/h). The low K-value for FP-2 signifies an exceptionally dense and electronically insulating SEI, effectively suppressing continued electrolyte reduction. The high K-value for FP-3 aligns with the expectation of a more porous SEI formed under high current density.
Electrochemical Impedance Spectroscopy (EIS) was conducted at 50% SOC. The Nyquist plots typically showed a semicircle in the high-medium frequency range (associated with charge transfer resistance, Rct, through the SEI) and a low-frequency Warburg tail (related to solid-state diffusion). We fitted the data to an equivalent circuit model R(QR)(QR)W to extract Rct. The charge transfer resistance represents the kinetic barrier for lithium ions to cross the SEI and engage in the redox reaction. The values were: FP-2: 12.5 mΩ, FP-1: 14.8 mΩ, FP-3: 18.2 mΩ, FP-4: 16.7 mΩ. The lowest Rct for FP-2 suggests its SEI has optimal ionic conductivity, likely due to a balanced organic/inorganic composition and good adhesion, facilitating rapid Li+ transport. This is crucial for high-rate performance.
Rate capability tests were performed by charging at 1C to 4.2V (CV cutoff 0.05C) and discharging at progressively higher rates from 1C to 12C. The capacity retention relative to the 1C discharge capacity is plotted. At moderate rates (≤6C), differences were minimal (all >95%). At ultra-high rates (8C-12C), the advantage of FP-2 became pronounced. At 12C, the retention was: FP-2: 87.5%, FP-1: 84.2%, FP-4: 81.0%, FP-3: 78.5%. This demonstrates that the robust SEI from the high-voltage shallow cycling protocol maintains structural integrity under high current stress, preventing excessive polarization and active material detachment. The rate capability can be linked to the internal resistance and diffusion limitations. The effective diffusion coefficient \( D_{\text{eff}} \) can be estimated from the low-frequency EIS Warburg region or from pulse tests. A higher \( D_{\text{eff}} \) correlates with better rate performance. The SEI from FP-2 appears to present a lower barrier for Li+ desolvation and entry into the anode.
High-temperature storage is a stringent test for SEI stability. Cells were fully charged and stored at 60°C for 28 days. Performance was assessed by capacity retention (capacity after storage / initial capacity), capacity recovery (capacity after a full recharge post-storage / initial capacity), thickness expansion, and DC internal resistance (DCR) increase. The results are consolidated in the table below.
| Protocol | Capacity Retention | Capacity Recovery | Thickness Expansion (Hot) | DCR Increase (%) |
|---|---|---|---|---|
| FP-1 | 86.3% | 91.7% | 3.4% | 4.5% |
| FP-2 | 87.5% | 92.2% | 2.4% | 3.9% |
| FP-3 | 84.7% | 88.5% | 3.5% | 11.1% |
| FP-4 | 85.9% | 91.2% | 3.5% | 2.3% |
FP-2 again showed the best overall storage stability, with the highest retention and recovery, low swelling, and a moderate DCR increase. The low thickness expansion suggests minimal gas generation from electrolyte decomposition, a sign of a stable SEI. FP-3 exhibited the largest DCR jump (11.1%), indicating significant SEI degradation and possible deposition of resistive decomposition products at the interface. Interestingly, FP-4 had the smallest DCR increase, likely because the higher-temperature formation produced an SEI with components that are more stable at elevated temperatures, but this came at the cost of slightly lower capacity recovery compared to FP-2.
Long-term cycle life at elevated temperature (45°C) under a 2C charge / 2C discharge regime further differentiated the protocols. The capacity retention over 300 cycles is plotted. The fade can be modeled using an empirical equation often applied to lithium-ion battery capacity fade:
$$ Q(n) = Q_0 – k \sqrt{n} $$
where \( Q(n) \) is capacity at cycle \( n \), \( Q_0 \) is initial capacity, and \( k \) is a fade rate constant. A lower \( k \) indicates better cycling stability. From the data, the fade rate constants were estimated: FP-2: \( k \approx 0.015 \), FP-1: \( k \approx 0.018 \), FP-4: \( k \approx 0.021 \), FP-3: \( k \approx 0.025 \). After 300 cycles, the capacity retention stood at: FP-2: 88.5%, FP-1: 86.0%, FP-4: 83.5%, FP-3: 80.2%. The superior cycle life of FP-2 is attributed to its SEI’s mechanical resilience. The shallow cycling during formation likely allowed the SiO particles to undergo an initial, controlled volume expansion, “pre-stressing” the SEI and forming a flexible, adhesive layer that could accommodate subsequent cycles without catastrophic fracture. This reduces the cumulative irreversible lithium loss per cycle.
The chemical composition of the SEI, inferred from performance data, can be conceptually linked to the formation parameters. A high-quality SEI should have a bilayer structure: a thin, dense inorganic inner layer (LiF, Li2O, Li2CO3) blocking electrons, and a thicker, polymeric organic outer layer (e.g., polycarbonates, oligoethers) allowing Li+ transport and providing flexibility. The formation current density \( j_{\text{form}} \) directly affects the nucleation rate \( N \) of SEI components. A classical nucleation and growth model suggests:
$$ N \propto \exp\left(-\frac{\Delta G^*}{kT}\right) $$
where \( \Delta G^* \) is the critical nucleation barrier, inversely proportional to overpotential \( \eta \). High \( j_{\text{form}} \) (FP-3) leads to high \( \eta \), low \( \Delta G^* \), and thus a high nucleation rate, resulting in many small nuclei and a fine-grained, possibly porous structure. Low to moderate \( j_{\text{form}} \) (FP-1, FP-2) allows fewer, larger nuclei to grow, forming a denser film. The addition of cycling in FP-2 may promote Ostwald ripening, where smaller, unstable SEI components dissolve and re-deposit on more stable regions, further consolidating the layer.
Temperature’s role is governed by the Arrhenius equation for reaction rate constants \( k_r \):
$$ k_r = A \exp\left(-\frac{E_a}{RT}\right) $$
Higher temperature (FP-4) increases all reaction rates, including desirable SEI-forming reactions and undesirable solvent co-intercalation or decomposition. This can lead to a thicker SEI with more organic species, which might be less stable long-term despite lower initial impedance.
In conclusion, this systematic investigation into formation processes for silicon oxide-graphite composite anode lithium-ion batteries demonstrates that protocol design must account for the distinct electrochemical and materials science principles governing each component. The high-voltage shallow cycling formation (FP-2) emerged as the optimal strategy. It successfully integrates kinetic and thermodynamic considerations: using low initial currents to form a dense SEI base, progressing to moderate currents for full lithiation, and incorporating intentional, shallow cycles in the high-voltage window where both graphite and silicon oxide are active. This approach engineers a solid-electrolyte interphase that is dense, ionically conductive, electronically insulating, mechanically flexible, and strongly adherent. Consequently, cells formed with this protocol exhibited outstanding performance across the board: low self-discharge (K-value), high first-cycle efficiency, low impedance, excellent rate capability, superior high-temperature storage resilience, and extended high-temperature cycle life. These findings provide a actionable framework for optimizing formation processes in advanced lithium-ion battery manufacturing, particularly for next-generation anodes incorporating silicon. Future work could involve in-situ spectroscopy to directly analyze the SEI composition and morphology resulting from each protocol, and further refinement of the cycling parameters (number of cycles, depth, voltage limits) for specific silicon oxide content and electrolyte formulations. The pursuit of higher energy density lithium-ion batteries continues to hinge on such precise control over interfacial chemistry from the very first cycle.
From an industrial perspective, the FP-2 protocol, while slightly more time-consuming due to the cycling steps, offers significant long-term benefits in cell performance and longevity, potentially reducing warranty costs and enhancing product reliability. The principles elucidated here—balancing kinetics, thermodynamics, and mechanical stress management during formation—are broadly applicable to the development of robust lithium-ion batteries with high-capacity alloying anodes. As the demand for electric vehicles and grid storage grows, optimizing such fundamental manufacturing processes will be key to delivering safe, durable, and high-performance energy storage solutions.