Lithium-ion battery electrode inspection using flash thermography

Nathan D Sharp, Purdue University

Abstract

Nonuniformity in lithium ion battery electrode thickness or composition can lead to reduced performance and longevity. Currently battery manufacturers have no way to quickly and accurately assess electrode quality during the manufacturing process. A finite element heat transfer model based on heat conduction equations has been developed to provide theoretical justification and insight. The model shows that a heat pulse to the back of a current collector will conduct through the electrode in such a way that spatial changes in thickness or material properties will have different transient temperature responses and that the response difference will be maximum around 3 to 10 ms after the flash occurs. Due to frequency limitations in the thermal camera microbolometers and effects from rastering, the raw temperature data was not able to accurately measure the transient response of the flash thermography signals. However, since the microbolometer time constant and rastering specifics are both known, an algorithm has been developed which uses the measured data to estimate the true response. A high end photon based thermal camera was used to validate the accuracy of the correction algorithm and has shown that the FLIR A325 camera with the correction algorithm can accurately measure the transient temperature response, but that it cannot measure the films with as much detail as the photon based camera due to reduced frequency capabilities. Therefore, this method will be sufficient for large film variability, but if a very detailed thermal picture is desired a more expensive camera will be required. Experiments were run to test the effectiveness of the flash thermography method for detecting several different types of defects. Gross defects such as contaminants, scratches, and bubbles were shown to be easily detectable. Thickness variation was also tested and shown to have a sensitivity of one percent change in temperature for one percent change in thickness. Thickness differences were shown to be detectable in at least as small as four percent thickness difference. Composition differences were also tested, looking at the difference in relative percentage of active material, carbon black, and PVDF. Not enough data was taken to quantify the sensitivities of composition changes, but testing was shown to be able to detect composition differences. Thermography testing also showed a wavelike thickness pattern occurring which has not previously been reported on battery electrodes. Comparison with a commercially purchased electrode showed that this phenomenon exists on the commercial electrode as well. Further testing needs to be conducted to determine the cause of this phenomenon, but it is hypothesized that is due either to a vibration in the coater blade or a nonlinear fluid interaction of the electrode slurry. Results and analysis show that flash thermography is a viable method to detect variability and defects in battery electrodes during the manufacturing process.

Degree

M.S.M.E.

Advisors

Adams, Purdue University.

Subject Area

Chemical engineering|Mechanical engineering

COinS