Polycrystalline cathode active materials in lithium-ion batteries have superior potential response. In order to achieve a longer battery life, it is necessary to prevent the polycrystalline particles from cracking during charge/discharge cycling. An investigation into the interfacial reactions, especially the side reactions at the grain boundary, is presented here to realize highly stable cathode-active polycrystalline particles. We propose that the formation of a resistive solid electrolyte interphase (SEI) by interfacial side reactions leads to volume expansion of the grain boundary, consequently causing the cracking of polycrystalline particles. An example of using grain boundary modification to prevent cracking during charge/discharge cycling is discussed. An epitaxial thin film electrode of (104)-oriented Li(Ni,Co,Mn)O2 cathode-active material with surface modification was prepared using pulsed laser deposition (PLD), in order to investigate how to protect the grain boundaries of cathode particle products. Several effective modifications were observed to suppress both resistive film formation in the grain boundary and the crystal structure destruction of cathode materials. These modifications prevented particle cracking in spite of the crystal volume change during repeated Li+ inter/de-intercalation. Cracking causes the internal resistance of the particles to increase, resulting in capacity fading of the battery. The cross section of deteriorated cathode electrodes shows a notable degree of cracking growth along the grain boundaries of the cathode active particles. Nano secondary ion mass spectroscopy (NanoSIMS) was carried out to confirm the elemental composition at the grain boundaries. The NanoSIMS analysis revealed that the cracks are filled with a typical resistive SEI, which contained species such as lithium alkyl carbonate and lithium fluoride. Although these components are known byproducts formed on the particle surface, it is notable that the reaction field extends into the grain boundaries. Boundary cracking is often discussed as a consequence of lattice expansion during electrochemical Li+ de-intercalation. However, lattice expansion also has the benefit of improving ionic conduction due to the expansion of Li+ diffusion path in the lattice. Therefore, it is important to develop polycrystalline particles that do not crack with permissible degree of volume change in the crystals. In order to study the efficient protection of the grain boundary from the resistive byproduct generation, epitaxial thin film electrodes oriented in the (104) crystal plane were prepared using PLD technique based on a previous work [1]. A protective film coating was then deposited on the surface of one thin electrode. A comparison of charge/discharge cycling tests for electrodes with and without protective coating shows a clear difference in the capacity and resistive film formation. Almost no resistive SEI was generated on the coated electrode, while thick SEI growth was observed on the pristine electrode, as shown in Figure 1. Moreover, the protective coating was effective in maintaining both the original electrochemical capacity and the crystal structure. Based on these results, another experiment was performed to modify the grain boundaries of the product level cathode-active particles with practical technique. The composition introduced in the grain boundaries was confirmed with NanoSIMS measurement. The particles with a modified grain boundary showed superior charge/discharge cycling performance. The cross section of the particles shown in Figure 2 is well consistent with the aforementioned PLD model experiment, where neither cracking nor SEI growth was observed along the particle grain boundaries. Thus, we have proven that epitaxial thin film electrodes are useful not only for theoretical studies but also for practical applications. The method described here can be used to suppress undesired interfacial reactions at the grain boundary in cathode active materials. [1] K. Sakamoto, M. Hirayama, N. Sonoyama, D. Mori, A. Yamada, K. Tamura, J. Mizuki, and R. Kanno, Chem. Mater. 21 (2009) 2632. Figure 1