E-Mobility Engineering 017 l ECE Doosan electric excavators dossier l In Conversation: Matt Faulks l Battery testing focus l Battery Show North America 2022 report l Ariel Hipercar digest l Cathode materials insight l Thermal management focus

The strength of the magnetic field was found to be dominated by the high current density at the input and output of the current tabs on the cell. The current is concentrated at the input and dies off away from these tabs. As the exact current density distribution is characteristic of the cell’s effective conductivity and cell geometry, any changes in the magnetic field distribution can be used to infer cell properties, such as the amount of lithium in an electrode, called the lithiation state, and how fast the lithium ions move through the cell. To investigate effects not captured by the model, such as changes in the SoC, the non-invasive test system can also measure magnetic fields changes that emerge at different charge stages. Starting with a battery with an open cell potential of around 3.7 V, the cell is cycled between charge states with a constant current of 10 A in stages of 1.3 Ah. This provides magnetic field changes between the cycles, corresponding to the SoC. This data can be used to refine simulation models. Larger magnetometer arrays will allow the simultaneous capture of data from across the battery and remove the need to move the FGA array across the material. This will allow the capture of current density images at higher rates of charge/discharge – situations in which equivalent circuit models often struggle to describe battery behaviour. Arrays with smaller sensor spacing, such as magnetometers based on nitrogen vacancy centres in diamond, could capture higher resolution images, potentially with a resolution of tens of nanometres. AI Machine learning (ML) has been used to predict battery health with 10 times higher accuracy than the current industry standard, which could aid in the development of safer and more reliable batteries for EVs. Focus | Battery testing 4D-STEM for non-invasive testing Adding a dimension to the various cell measurements is helping scanning transmission electron microscopy (4D-STEM), coupled with electron energy loss spectroscopy (EELS) and data mining to explore the structure of cubic spinel manganese dioxide nanoparticles in a cathode. The 4D aspect of STEM uses an electron detector to capture a convergent beam electron diffraction (CBED) pattern at each scan location. This captures a 2D image at each scan point as the beam moves across the two dimensions of the sample. This has been enabled by developments in the electron detectors and improvements in processing. This technique is more often used for examining ceramic materials, but when used for electrode materials it gives a tenfold increase in the resolution of the structure of nanoparticles over X-ray systems. During the operation of rechargeable ion batteries, ions diffuse in and out of the electrodes, causing mechanical strain and sometimes cracking failures. The 4D-STEM technique captures the strain-caused nanoscale domains inside battery materials for the first time. This is critical for mapping otherwise inaccessible variations of crystallinity and domain orientations inside the materials. Adding complex data analysis, or data mining, to identify structures in the materials show a pattern of nucleation, growth and coalescence inside the batteries as the strained nanoscale domains develop. This non-invasive test approach also shows that the electrolytes have a substantial impact on the transformation of the microstructure. Large strain gradients build up from the development of phase domains across their boundaries with high impact on the chemical diffusion coefficient by a factor of ten or more. IN 4D-STEM the data is collected in the same way as standard transmission electron microscopy, but with different detectors for faster measurements. Rather than phosphorescent scintillators paired with a charge-coupled device, 4D-STEM uses a pixelated electron detector located at the back focal plane that collects the CBED pattern at each scan location. This sensor can be a monolithic active pixel sensor, which is a CMOS pixel array with a doped epitaxial surface layer that converts high-energy electrons into many lower energy ones that travel down to the detector. This gives a pixel resolution of up to 4000 x 4000, and enables the detection of individual electronics. An alternative sensor is a pixel array detector. This consists of a photodiode bump bonded to an integrated circuit, where each solder bump represents a single pixel on the detector. This type of detector has a smaller number of pixels, typically 128 x 128, but can achieve a much higher dynamic range of the order of 32 bits and high-speed detection of 1 ms per pixel. Using a hollow detector with a hole in the middle allows transmitted electrons to be passed to an EELS detector while scanning. This allows for the simultaneous collection of chemical spectra information and structure information. The computation aspect is also important. The detectors generate vast amounts of data, as much as 100 Tbytes per scan. ML algorithms can dig through the data to identify patterns that give more insight into the structures inside a battery. 40 January/February 2023 | E-Mobility Engineering