Lithium ion batteries are one of the most promising energy storage systems that have been widely used in portable electronic devices and electric vehicles. Nevertheless, advanced automotive electrification requires greater capacity, longer cycle life and reduced cost. With respect to cycle life, the mechanisms of capacity degradation are not yet to be fully understood. What is known is that solid electrolyte interface (SEI) formation ^[1] and structural distortion of active materials adversely impacts battery performance.^[2,3] Furthermore, next generation negative electrodes are known to suffer huge volume changes during Li⁺ intercalation/deintercalation upon cycling, which induce local stresses and lead to microstructure degradation of the electrode.^[4] While the smaller volume changes of the cathode materials are less well recognized, they may also turn out to be significant for battery performance and life.^[5]

In the present work, the structural evolution of the LiNi_xMn_yCo_1-x-yO₂ (NMC) and Li_1.2Ni_0.13Mn_0.54Co_0.13O₂ (HE-NMC) cathodes in lithium-ion batteries following electrochemical cycling were studied three-dimensionally.^[6] The cathodes were successfully reconstructed in 3D using state–of-the-art focus ion beam-scanning electron microscope (FIB-SEM) serial sectioning and imaging techniques. The acquired 2D images were digitally segmenting into three distinct phases before reconstruction. In this study, a segmentation algorithm was developed to apply refined threshold values and gradient analysis to the 2D images, thereby avoiding the necessity of epoxy infiltration to the sample ^[7] and protect the integrity of the small features of the phases. As shown in Figures 1a-1b, each of the three phases: active material (AM), carbon-doped binder (CB), and pore spaces (which are filled with electrolyte during cycling) were clearly resolved after the image processing step.

In order to evaluate the evolution of the relative positions of the important phases within the cathode, the segmented 3D structure was quantified with a “neighbor counting” method, by which the connectivity between CB and the AM was analyzed for cathodes that underwent different number of cycles. This approach provides statistical analysis-based assessment of the degree of connectivity between the surface of active particles and the binder matrix as a function of depth into the cathode (i.e., in the direction perpendicular to the separator and current collector). The probability of carbon-doped binder being adjacent to the active particles (PBA) of cycled NMC cathode is shown in Figure 1c. Each data point represents the probability of CB being in direct contact with the active particle surface of each slice perpendicular to the depth of the 3D structure in units of pixels, each pixel corresponding to the slice thickness (35 nm), as shown in the schematic diagram in Figure 1c.

It is shown that the average PBA of the pristine sample decreases from 0.54 to 0.48 after 20 cycles, and further drops to 0.43 after 50 cycles. The value of the average PBA, e.g., 0.54, means that 54% of the AM surface is in contact with the CB phase within the 3D volume. The results therefore indicate that the connectivity between AM and CB decreases with the increased number of cycles, where a cycling-induced detachment between the two phases was observed. Similar detachment phenomenon is also observed in the HE-NMC cathode after 50 cycles, for which a 22.5% loss of AM and CB contact was detected. Since the CB phase contains the carbon black additive for the enhancement of electrical conductivity of the cathode, our results thus suggest that the detachment between the conductive CB phase and the active particles will reduce the efficiency of electron transport from the active particles (where Li⁺ intercalation/deintercalation happens) to the current collector. The loss of electrical conductivity of the cycled cathode will have an adverse effect on the cyclic performance as a consequence of the detachment between the conductive binder matrix and active particles.

[1] P. Verma, P. Maire, P. Novak, Electrochem. Acta. 2010, 55, 6332-6341.

[2] F. Lin, I.M. Markus, D. Nordlund, et al. Nat. Commun. 2014, 5, 3529.

[3] A. Boulineau, L Simonin, J.F. Colin, et al. Nano Lett. 2013, 13, 3857-3863.

[4] D. Larcher, S. Beattie, M. Morcrette, et al. J. Mater. Chem. 2007, 17, 3759–3772.

[5] D.S. Eastwood, V. Yufit, J. Gelb, et al. Adv. Energy Mater. 2014, 4, 1300506.

[6] H. Liu, J. M. Foster, A. Gully, et al. J. Power Sources. 2016, 306, 300-308.

[7] Z. Liu, J. Scott Cronin, Y.K. Chen-Wiegart, et al. J. Power Sources. 2013, 227, 267–274.