In Situ tem Cycling of Semi-Solid State Micro-Battery in Liquid Electrolyte

Abstract

All solid state Li-ion micro-battery is a promising candidate to power miniaturized sensors for Internet of things (IOT) and other electronic devices. In recent times, the spinel LiMn1.5Ni0.5O4 (LNMO) has demonstrated as a potential positive electrode material for Li-ion thin film batteries offering a theoretical capacity of 147 mAh/g (65 µAh/cm2/µm for a bulk density of 4.47 g/cm3) and operates up to now at the highest potential (around 4.7 V vs. Li+/Li) [1]. In this work, we report our first successful in situ TEM attempts to observe the morphological, structural and interfacial changes in the positive electrode layer of FIB prepared sample which undergo after cycling using liquid electrolyte. More precisely we compared the morphological and structural evolution between a pristine and cycled microbattery by 4D STEM-ASTAR technique to highlight the key information to improve the deposition conditions that will enhance the reliability and production quality of such micro power devices. The in situ cycling in liquid TEM has given this opportunity to study the battery electrode materials so as to spot the slightest modifications of the materials resulting in important advances in knowledge on electrochemical energy storage [2-3].Here, our approach is based on the cycling a FIB lamella sample inside the TEM using liquid-electrochemical TEM holder with conventional liquid electrolyte (1M LiClO4, EC: DMC 1:1). The cross-section image of as prepared FIB sample with homogeneous deposition of distinctive layers of different thickness [from bottom to top-Si (0.385mm)/ Al2O3 (100nm)/ Pt (630nm) / LNMO (400nm)] is shown in figure 1a. The Pt current collector of FIB lamella sample is connected to the Pt working electrode on the e-chips used for TEM study (figure 1b). First, using FIB preparation technique, we sliced a full 2-D “thin film micro-battery” making it as thin to observe/analyse under TEM. Then, we modified the FIB lamellar design using FIB-SEM tool to get good electrical contact and reduced polarisation. Several technological problems have to be overcome in the process. For the instance, it is mandatory to obtain a good electrical contact between the Pt working electrode of e-chip and Pt current collector of FIB lamellar, which is later achieved by depositing extra Pt between the two contacts forming a platinum bridge. A 4-Probe electrical conductivity performed locally confirms the good electrical contact between the as prepared FIB lamella sample and Pt working electrode of e-chip. The cross-section bright field TEM image of a final modified version of FIB lamella sample used in the study is shown in figure 1d.The FIB lamella sample is then later cycled inside the liquid electrochemical TEM holder (fig 1c) in potential window of 4.1 V-4.8 V vs. Li+/Li. The flow of the electrolyte (LiClO4 EC:DMC 1:1) inside the TEM holder was further controlled by microfluidic controller with the flow rate of 2 µL/min. CV was recorded at a sweep rate of 0.1mV/s and two plateaus at 4.4 V and 4.6 V was observed corresponding to Ni2+/3+ and Ni3+/4+ oxidation respectively (Fig 1f). The basic redox steps observed during the charge are same as observed in the case of cycling bulk 2D thin film in a homemade flat cell.The comparison between the cycled and pristine microbattery sliced by FIB and observed by TEM allowed us to clearly demonstrate the formation of cracks inside the LMNO layer, loss of contact between the LMNO layer and the platinum current collector, as well as the agglomeration of the organic compounds produced due to the electrolyte decomposition. Moreover, 4D STEM-ASTAR technique provided us with the crucial information regarding the grain size reduction from 20 nm in pristine to 12 nm in cycled sample, confirming the continuous electrode-electrolyte reaction happening over the cycling. Also, the decrease in crystallinity and increase in the amorphization of the LMNO grains by 38 % in the cycled sample compared to pristine as shown in figure 1e and g clearly prove the fast capacity fading phenomenon observed in bulk microbattery. Furthermore, a thickness of ~20nm along the platinum layer, a (111) preferred orientation is observed exhibiting the epitaxial effect of LMNO on platinum layer which has been further supported by the precision electron diffraction (PED) recorded on both LMNO and platinum grains.