Multi-Method Li Plating Characterization of a Commercial 26 Ah Li-Ion Pouch-Cell

1. Problem Statement

Lithium (Li) plating on graphite anodes is a critical degradation mechanism in Li-ion batteries, leading to capacity loss (due to "dead" Li) and severe safety risks (dendrite-induced short circuits and thermal runaway). While numerous methods exist to study Li plating in laboratory cells, investigations on commercial high-energy cells often rely solely on electrochemical methods, which lack direct elemental confirmation. There is a lack of comprehensive, comparative studies that validate multiple detection techniques (electrochemical, optical, microscopic, and spectroscopic) on the same commercial cell to ensure data comparability and reliability.

2. Methodology

The study employed a "round-robin" approach across multiple research laboratories to characterize a single commercial A123 26 Ah pouch cell (NMC 111 cathode, graphite anode, no silicon).

• Experimental Setup:

  • Cell Cycling: The cell was subjected to fast charging at 1 C rate and -5 °C to induce Li plating. A control cycle was performed at 25 °C.
  • Electrochemical Analysis: Stripping Differential Voltage Analysis (DVA) was used to detect reversible Li plating via characteristic voltage plateaus during discharge.
  • Cell Opening: The cell was opened immediately (<10 min rest) in an Argon-filled glovebox to prevent Li re-intercalation or oxidation.
  • Sample Distribution: Anode sheets were scanned, cut, and distributed to various institutes for independent analysis.

• Multi-Method Characterization:

  • Optical & Imaging:
    • Flatbed Scanner: High-resolution (600 DPI) scanning of full anode sheets, processed with Deep Learning (U-Net) to create heatmaps of plating intensity.
    • Light & Laser Scanning Microscopy: Visualized surface morphology and 3D height profiles of deposits.
    • X-ray Microscopy (XRM): Provided 3D internal structural views.
  • Microscopy & Elemental Analysis:
    • SEM/EDX: Used standard and windowless EDX detectors to directly detect Lithium (Li Kα at ~54 eV) and map elemental distribution.
    • FIB-SEM: Created cross-sections to analyze depth profiles and subsurface structures.
    • Colocalized Analysis: Combined SEM with Raman Spectroscopy and Soft X-ray Spectroscopy (SXES) to correlate morphology with chemical composition.
  • Spectroscopic & Bulk Analysis:
    • NMR (7Li): Identified metallic Li vs. intercalated Li based on chemical shifts (e.g., ~265 ppm for metallic Li).
    • EPR: Detected unpaired electrons in metallic Li, distinguishing between mossy, dendritic, and intercalated states based on line shape.
    • GD-OES & NDP: Provided depth profiling of Li concentration from the surface into the bulk material.
    • ICP-OES: Quantified total Li content after dissolving samples.

3. Key Contributions

• First Comprehensive Round-Robin Study: This is the first study to apply a wide spectrum of detection methods (electrical, optical, microscopic, spectroscopic) to the same commercial cell, ensuring high comparability of results.
• Validation of Direct Li Detection: Demonstrated the capability of windowless EDX and SXES to directly detect metallic Li on graphite, overcoming the limitations of standard detectors.
• Correlative Imaging: Successfully combined optical imaging (for spatial distribution) with spectroscopic methods (for chemical confirmation) to create a holistic view of Li plating.
• Deep Learning Integration: Utilized computer vision and U-Net models to automate the segmentation and quantification of Li plating intensity across entire electrode sheets.
• Methodological Benchmarking: Provided a detailed comparison of the advantages, limitations, spatial resolution, and measurement duration for each technique.

4. Key Results

• Electrochemical Evidence: DVA confirmed Li plating with a characteristic stripping plateau at -5 °C/1 C, indicating ~1.5% of nominal capacity was plated.
• Spatial Distribution:
  • Plating was heterogeneous, concentrated heavily near the anode edges and specifically on the side of the cathode tab.
  • Outer sheets in the stack showed significantly more plating than middle sheets, attributed to temperature gradients (outer sheets cooled faster, leading to higher overpotentials).
• Morphology:
  • Deposits appeared as silver/white needle-like structures (dendrites) and mossy islands.
  • Needle lengths were measured at 10–15 μm, growing from the graphite surface toward the separator.
• Elemental Confirmation:
  • EDX/SXES: Directly detected Li Kα peaks (54 eV) in needle structures, confirming >70 at% metallic Li.
  • NMR: Showed a small signal (~0.3% of total area) at ~265 ppm, characteristic of metallic Li, alongside intercalated Li signals.
  • EPR: Detected a very narrow signal (ΔBpp = 0.003 mT) specific to metallic Li, confirming its presence even when NMR signals were subtle.
• Depth Profiling:
  • NDP and GD-OES confirmed that Li plating is primarily a surface phenomenon (top 2–4 μm), with concentration dropping to bulk levels at depths >8 μm.
  • NDP revealed higher Li concentrations in the corners and edges of the anode sheets compared to the center.
• Quantification Discrepancies: ICP-OES showed similar total Li in plated and non-plated regions, highlighting its inability to distinguish between intercalated and metallic Li without spatial resolution.

5. Significance and Recommendations

• Method Selection Guide: The paper establishes a hierarchy for Li plating detection:
  • Optical methods (Scanner/Microscopy) are recommended for rapid, large-area screening due to speed and ease of use.
  • Spectroscopic methods (SEM-EDX, EPR, NMR) are essential for reference samples to chemically confirm the presence of metallic Li and distinguish it from SEI or intercalated Li.
• Safety and Reliability: The study validates that DVA stripping peaks are reliable indicators of plating onset, but post-mortem analysis is required to understand the physical nature of the deposits.
• Future Applications: The validated data and multi-modal approach provide a robust dataset for simulating Li plating, allowing researchers to model spatial distributions and chemical reactions more accurately.
• Best Practice: The authors recommend a workflow where optical methods identify regions of interest, which are then verified by spectroscopic methods to ensure accurate quantification and material identification.

In conclusion, this work provides a definitive "toolkit" for characterizing Li plating in commercial batteries, bridging the gap between electrochemical indicators and physical/chemical reality through rigorous, multi-laboratory validation.