Computed Tomography Scanning of Battery Cells

Computed Tomography (CT) Scanning of Battery Cells has been used a lot in research and as an offline manufacturing quality check.

CT scan requirements [1]:

Able to detect metal particles 50 μm x 50 μm x 20 μm
Scan resolution of 70 μm minimum in the axial direction
Scan resolution of 50 μm minimum in the radial direction

The following defects list is from reference 1 and relates to cylindrical cells. However, with some small changes this list can be applied to all formats of cells with some slight changes:

Improper seating or pinching of the plastic seal ring
Improper routing of the positive tab from the current interrupt device (CID) to the jellyroll (JR).
Excessive telescoping of the JR to point where anode current collector edge overlap relative to the cathode edge is non-existent in any of the JR windings either at the top or bottom. See Figure 1 for an example of nominal vs failing.
Excessive bending or deformation of the anode current collector overlap in the top or bottom of the JR. This usually comes with defect.
High density anomalies outside the JR could be indications of weld splatter.
Internal crimp seal corrosion pitting on the inside of the can next to the seal ring.
High density anomalies inside the JR could be indications of electrode wrinkles, foreign object debris (FOD) or native object debris (NOD).
Anomalous can wall thin spots.

Using CT Scanning to Detect Battery Defects

Glimpse take us into the cell and demonstrate what defects can be found using these tools.

Analysis Workflows for XCT Data

A set of open Jupyter notebooks and machine learning (ML) models that demonstrate analysis workflows for battery XCT data. In version 1.0.3 the workflows target three critical QA tasks relevant to Li-ion cylindrical cells with the fourth notebook demonstrating relevant image segmentation approaches.