Review of ultrasonic methods for monitoring, damage detection, and processing of lithium-ion batteries throughout their life-cycle

Technical Summary: Review of Ultrasonic Methods for Monitoring, Damage Detection, and Processing of Lithium-Ion Batteries

Problem Statement
Lithium-ion batteries (LIBs) are the dominant energy storage technology for consumer electronics, electric vehicles (EVs), and grid applications. However, the rapid scaling of LIB adoption has exposed a critical gap in understanding their internal behavior throughout their life cycle. Current battery management systems (BMS) rely primarily on electrical and temperature measurements to estimate the State of Charge (SOC) and State of Health (SOH). These methods provide limited insight into complex electrochemical processes and mechanical changes, such as electrode expansion, gas generation, or internal short circuits, which are precursors to thermal runaway and safety failures. There is an urgent need for non-destructive testing (NDT) methods that can probe the internal mechanical and structural integrity of LIBs to enhance safety, predict lifespan, and facilitate second-life applications and recycling.

Methodology
This paper provides a comprehensive review of ultrasonic testing (UT) methods applied to LIBs, synthesizing research from the past decade. The review is structured around three primary pillars:

1. Fundamentals: It establishes the physics of wave propagation in the heterogeneous, layered, and poroelastic media of LIBs. This includes the application of Biot's theory to model fast and slow longitudinal waves in porous electrodes and separators, as well as the analysis of guided waves (Lamb waves) in plate-like structures.
2. State Estimation and Damage Detection: The authors categorize experimental approaches into bulk wave and guided wave methods. They analyze how ultrasonic metrics—specifically Time-of-Flight (TOF), Signal Amplitude (SA), and attenuation—correlate with SOC, SOH, and specific degradation mechanisms (e.g., lithium plating, SEI growth, electrolyte dry-out, and gas formation).
3. Modeling and Data-Driven Approaches: The review evaluates physics-based models (transfer matrix methods, finite element analysis, effective medium theories) and data-driven models (machine learning, neural networks) used to interpret ultrasonic signals. It highlights the challenges in linking acoustic observables to specific physical phenomena due to the complex coupling of electrochemical and mechanical behaviors.

The review also examines the application of UT across the entire battery life cycle, including manufacturing (electrode drying, calendaring), conditioning (SEI formation), usage (real-time monitoring), second-life screening, and recycling.

Key Contributions and Results

• Physics of Wave Propagation: The paper clarifies that LIBs function as layered poroelastic media. It details how the intercalation of lithium ions causes volume changes (e.g., ~13% expansion in graphite anodes) and modulus variations, which alter ultrasonic wave speeds and attenuation. It distinguishes between bulk waves (sensitive to through-thickness properties) and guided waves (sensitive to in-plane properties and specific modes like $A_0$ and $S_0$).
• State of Charge (SOC) and Health (SOH) Correlations:
  • SOC: A consistent trend observed in bulk wave studies is that TOF decreases and SA increases with increasing SOC, primarily driven by the mechanical stiffening and volume changes of the graphite anode. However, guided wave studies show conflicting trends (e.g., some report TOF increases with SOC), attributed to mode selection and frequency-dependent dispersion.
  • SOH: Aging generally leads to an increase in TOF (due to electrode thickening and mechanical softening) and variable trends in SA. The paper notes that "knee" points in capacity fade can be detected using joint time-frequency metrics.
• Damage Detection: UT has demonstrated high sensitivity to specific failure modes:
  • Gas Formation: Causes significant signal attenuation and loss of transmission due to acoustic impedance mismatch.
  • Lithium Plating: Detected via changes in signal amplitude and time-of-flight, particularly at low temperatures and high C-rates.
  • Thermal Abuse: UT can detect early signs of thermal runaway through rapid changes in TOF and signal loss prior to catastrophic failure.
• Modeling Insights: The review highlights that while physics-based models (using Biot's theory and transfer matrices) can predict band gaps and dispersion curves, they often lack precise material property data for specific cell chemistries and states. Data-driven models show high accuracy in SOC/SOH estimation but suffer from poor generalizability across different cell types and lack physical interpretability.
• Life Cycle Applications:
  • Manufacturing: UT can monitor electrode drying dynamics, coating uniformity, and electrolyte wetting.
  • Second-Life: UT offers a rapid, non-destructive method to screen retired EV batteries for internal defects (gas, delamination) that electrochemical testing might miss.
  • Recycling: UT can assess the state of spent cells before disassembly to prevent thermal runaway and sort cells based on internal damage.
  • Active Control: Emerging research suggests ultrasound can actively influence battery performance, such as enhancing ion transport via acoustic streaming or modifying SEI layers.

Significance and Outlook
The paper asserts that ultrasonic testing is a promising, non-invasive tool capable of providing unique mechanical insights into LIBs that complement traditional electrical monitoring. Its significance lies in its potential to reduce BMS weight and complexity, improve safety by detecting latent defects early, and enable the circular economy through efficient second-life screening and recycling.

However, the authors emphasize that the field is still in its infancy. Major challenges remain, including:

• Generalizability: Current methods often lack the ability to generalize across different cell chemistries, form factors (pouch, cylindrical, prismatic), and manufacturing batches.
• Material Property Data: There is a critical lack of standardized, experimentally validated mechanical property data for cell constituents under various operating conditions (SOC, temperature, aging), which hinders the development of accurate predictive models.
• Engineering Implementation: Practical integration into commercial BMS requires robust, cost-effective transducer designs, reliable coupling methods, and real-time data processing capabilities that can withstand the harsh environments of EVs.
• Interpretability: Bridging the gap between observed ultrasonic signals and specific physical degradation mechanisms requires improved physics-based modeling and multi-modal validation (e.g., combining UT with X-ray CT).

The review concludes that while significant progress has been made, realizing the full potential of UT in LIBs requires concerted efforts in standardizing measurement protocols, creating open-access material property databases, and developing hybrid modeling approaches that integrate electrochemistry with wave physics.