Assessing the Accuracy of Lithium Contents Determined by Combined Quantitative Backscattered Electron and X-ray Energy Dispersive Spectroscopy Analysis

Le résumé fourni par la source

Lithium- (Li-) based materials and products have gained significant commercial importance due to their widespread utilization in applications such as battery technologies and aerospace alloys. However, to understand and optimize properties of these materials, it is highly desirable to determine the elemental composition at the microscale quantitatively. However, commonly used elemental analysis techniques in the scanning electron microscope (SEM), like X-ray energy dispersive spectroscopy (EDS) and secondary ion mass spectroscopy (SIMS), are unsuitable for quantitative analysis of lithium. However recently, an alternative approach for quantitative evaluation of the Li content has been described using a composition-by-difference technique based on comparison of quantification (of non-lithium elements) by EDS and quantitative backscattered electron (qBSE) microscopy. Researchers have demonstrated an accuracy significantly better than 1 wt. % in a range of material systems including LiAlMg alloys [1], stoichiometric compounds [2] and cathode materials e.g., lithium nickel manganese cobalt oxide (NMC) [3].In this study, we used the Cipher® system (Gatan Inc.) attached to an FE-SEM to investigate the precision of lithium measurements performed using the EDS/qBSE composition by difference method. We performed a systematic investigation to understand the influence of the SEM settings and other parameters to identify a ‘Goldilocks’ set of experimental conditions including accelerating voltage, beam current, dwell time and working distance that minimized the variance in qBSE measurements. The precision of a lithium measurement that can be achieved using the Cipher system was investigated by analyzing the contribution of crystal orientation on the BSE signal. qBSE measurements were collected from 150 LiF crystals oriented randomly and cross sectioned with their surface normal to the electron beam direction. Optimized experimental conditions were determined achieving a variance in atomic number smaller than ± 0.11 atomic numbers (± 1.4 %); equivalent to a lithium fraction of 48.8 ± 2.4 at. %, Figure 1. Moreover, we investigated the precision to which the lithium content can be analyzed using the Cipher system at the primary particle level in an NMC 811 powder that is widely used as a cathode material in li-ion batteries. An NMC 811 powder of nominal lithium content 25.0 ± 4.1 at. % was prepared for analysis by embedding in epoxy before being cross sectioned by broad beam argon milling at -50 °C using the PECS™ II system (Gatan Inc.). Approximately 100 secondary particles, ranging from approximately 5 to 20 µm in diameter were analyzed and the mean lithium content of secondary particles determined to be 23.8 ± 3.9 at. %, consistent with nominal values. The secondary particles consisted of several hundred smaller primary particles, typically 50 - 1,000 nm in size which could be visualized clearly in a BSE image captured at an accelerating voltage of 3 kV, Figure 2. However, when the same secondary particle was imaged at 10 kV—conditions more suitable for qBSE analysis—the contrast between primary grains was greatly reduced and furthermore, was not consistent between the two analysis conditions. The variance in the qBSE value of primary grains was ± 2.2 %, marginally higher than in the LiF case. To determine whether this can be a real variation in the lithium content or the limit of the precision of the technique, a correlative study of qBSE signal versus crystal orientation collected using the Clarity direct detection EBSD system (EDAX LLC) will be presented. This accuracy and precision of results demonstrated provide critical evidence that analysis of lithium by combined qBSE/EDS provides a viable method to study lithium distribution at the microscale and the potential to study migration of lithium during the charge-discharge cycle in battery cells. Lithium fraction of lithium fluoride particles as determined using the Cipher system. Backscattered electron images of a (lithium) nickel manganese cobalt oxide particle. (Left) High contrast image revealing <10 nm primary particles; captured at low accelerating voltage (3 kV) to enhance contribution of channeling contrast. (Right) Mean atomic number (Z) image; captured at moderate accelerating voltage (10 kV); BSE signal proportional to mean atomic number.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.