High-Resolution Atomic Absorption Spectrometry Combined With Machine Learning Data Processing for Isotope Amount Ratio Analysis of Lithium

Abstract

An alternative method for lithium isotope amount ratio analysis based on a combination of high-resolution atomic absorption spectrometry and spectral data analysis by machine learning (ML) is proposed herein. It is based on the well-known isotope shift of approximately 15 pm for the electronic transition 2 2 P←2 2 S at around the wavelength of 670.8 nm, which can be measured by state-of-the-art high-resolution continuum source graphite furnace atomic absorption spectrometry. For isotope amount ratio analysis, a scalable tree boosting ML algorithm (XGBoost) was employed and calibrated using a set of samples with 6 Li isotope amount fractions ranging from 0.06 to 0.99 mol mol −1 , previously determined by multi-collector inductively coupled plasma mass spectrometry (MC-ICP-MS). The calibration ML model was validated with two certified reference materials (LSVEC and IRMM-016). The procedure was applied to the isotope amount ratio determination of a set of stock chemicals (Li 2 CO 3 , LiNO 3 , LiCl, and LiOH) and a BAM candidate reference material, that is, LiNi 1/3 Mn 1/3 Co 1/3 O 2 (NMC111) cathode material. The results of these determinations were compared with those obtained by MC-ICP-MS and found to be metrologically comparable and compatible. The residual bias was −1.8‰ and the precision obtained ranged from 1.9‰ to 6.2‰. This precision was sufficient to resolve naturally occurring variations, as demonstrated for samples ranging from approximately −3‰ to +15‰. To assess its suitability to technical applications, the NMC111 cathode candidate reference material was analyzed using high-resolution continuum source molecular absorption spectrometry with and without matrix purification. The results obtained were metrologically compatible with each other. File list (2) download file view on ChemRxiv Manuscript_Preprint_20210115.pdf (703.18 KiB) download file view on ChemRxiv ESI_Manuscript_20210115.pdf (210.73 KiB)High-resolution atomic absorption spectrometry combined with machine learning data processing for isotope amount ratio analysis of lithium

Development of a Portable Method for Serum Lithium Measurement Based on Low-Cost Miniaturized Ultrasonic Nebulization Coupled with Atmospheric-Pressure Air-Sustained Discharge

“…Among the reported methods for inorganic elements in serum, some studies used nitric acid for digestion, 23 while some researchers indicated that Li in serum does not participate in the synthesis of proteins and therefore, digestion was needless. 14,17,18 To confirm this in our study, both highpressure bomb digestion (HBD) and direct dilution (DD) with deionized water were evaluated.…”

Determination of lithium in human serum by isotope dilution atomic absorption spectrometry

“…The Li characteristic mass (m c ) which represents 1% transmittance was found to be 0.6 pg. The ID analysis was based on the isotope shift of the electronic transition 2 2 P ← 2 2 S, which exhibits two spin-orbit components for each isotope, as we previously described [21]. The nonlinear fit parameters were optimized using the spectra of 7 Li-and 6 Li-enriched materials.…”

“…Since the four peaks reflect only two doublet transitions of the two isotopes, the ratio between two respective areas (e.g., A 1 and A 2 ) is constant. Namely, the spectral displacement between c 1 and c 3 is the isotopic shift of 15.80 pm between 7 Li and 6 Li [21]. The distance between c 1 and c 2 or c 3 and c 4 is the spin-orbit splitting of 15.08 pm.…”

“…The wavelength of the instrument was centered at 670.7845 nm. The optimized measurement conditions provided in our recent work were applied [21]. However, owing to the low vapor pressure of La, the atomization temperature was increased to 2500°C.…”

“…The same approach was applied for the trace analysis of Br and Ca [18,19]. Additionally, accurate isotope ratios, which are the core of ID-MS applications, can be achieved by HR-CS-MAS and HR-CS-AAS using modern methods for data analysis like partial least square regression and machine learning, as was demonstrated for B and Li, respectively [20,21]. In this latter work on Li isotope ratio analysis, the electronic transition 2 2 P ← 2 2 S was explored.…”