Lithium-ion batteries (LIBs) have profoundly advanced the development of electric vehicles (EV). However, one of the remaining bottlenecks in their widespread deployment is the long charging time typically required for commercial LIBs. There is a global push towards extreme fast charging (XFC) of EV batteries to reduce their charging times to 10-15 minutes.1 However, XFC causes severe degradation in the electrochemical performance of LIBs, which is mainly attributed to irreversible Li plating on the surface of the graphite anode rather than intercalation of Li ions inside the graphite as the battery is fast-charged.2 Since Li plating has been identified as one of the major XFC degradation mechanisms Bs, understanding its origin and spatial heterogeneity across three dimensions (3D) on porous graphite anodes as a function of time necessitates a non-destructive characterization platform to differentiate between graphite and Li.Neutron-based imaging provides the sensitivity to differentiate graphite from Li due to the large difference in their total neutron cross-sections (Li:C ~ 73:11 barns). However, high energy X-rays enable penetrating the metallic battery components. e.g., Cu, Al. We tested the feasibility of simultaneous neutron and X-ray tomography (NeXT) on the BT-2 imaging beamline at National Institute of Standards and Technology Center for Neutron Research.4 Our spatial resolution for both X-rays and neutrons was ~15-20 μm. For ex-situ Li plating detection, we imaged 1) pristine, and 2) a cycled graphite anode.5 For cycled anode, we disassembled the battery pouch cells in their fully discharged condition and harvested the anode after 450 cycles of 9C charging and C/2 discharging. For data analysis, we used 2D bivariate histogram phase segmentation to differentiate different battery materials such as Cu, graphite, etc. However, low X-ray energy (40keV) combined with the flat sample geometry of the graphite strips made it challenging to remove beam hardening artifacts from copper and, therefore, to segment lithium from graphite. In addition, to determine NeXT feasibility for in-situ imaging, we also characterized a fully assembled, uncycled single-layer battery pouch cell. However, the high neutron attenuation by the hydrogen in EC: EMC (3:7) electrolyte and hydrogenous polymeric materials in the Celgard separator made it particularly difficult to segment graphite anode from the separator and positively identify interfaces.In this presentation, we will outline the design and characterization of a neutron-friendly LIB coin cell for extreme fast-charging. Our design is based on a standard 2032 format coin cell with custom modifications to address the dual criteria of electrochemistry and neutron imaging, i.e., good neutron transmission and relative contrast from a functional LIB at 6-C charging rate. We will present three advantages of coin cell geometry: 1) it provides more uniform attenuation throughout the entire tomography scan thus, reducing artifacts; 2) multiple coin cells can be stacked on top of each other for high throughput battery imaging; 3) it reduces the amount of expensive deuterated materials needed while developing the optimum cell configurations. In addition, we will address the need of a deuterated EC: DMC electrolyte in a fluorinated separator for the neutron-friendly battery design (Figure 1). Since the total neutron cross-section of deuterium and fluorine are on the orders of magnitude lower than that of hydrogen, we anticipate that these modifications will increase neutron transmission through the battery, thus improving signal-to-noise ratio and reducing acquisition time especially at low neutron flux facilities. Secondly, they will result in a cleaner electrolyte-separator-anode interface for the detection of plated lithium because of absence of hydrogen. Finally, we will show how these modifications affect the electrochemical performance of our LIB coin cells to determine the onset of Li plating at 6-C charging rate. i-e., capacity fade vs. number of cycles.
Read full abstract