LIBS Info: Element Analysis

Title Authors Material Detector Spectrometer Software
A novel method to extract important features from laser induced breakdown spectroscopy data: application to determine heavy metals in mulberries Di Wu, Yong He, Jingyu Wang, Liang Yang, Liuwei Meng, Lingxia Huang Organic Andor iStar ICCD Mechelle 5000 Solis
Laser: Nd:YAG
Gate Delay: 4.000us
Gate Width: 16.000us
The authors demonstrate a novel machine learning analysis technique for Cu and Cr in Mulberries. Samples were dried, powdered, sieved and then pressed into pellets prior to analysis. A key outcome was that the analytical performance was significantly improved by not using the full spectra, but by a subset of the data.
Element Detection Limit (ppm) Wavelength (nm) Other Wavelengths (nm) Calibration Method Calibration Range (ppm) Notes

Element RMSE (ppm) Wavelength (nm) Calibration Method Notes
Cr 211.0000 ppm 425.4600 Univariate Calibration Univariate calibration for comparison with ML result
Cr 92.0000 ppm -10.0000 Machine Learning SVM result, following extensive pixel/spectra feature reduction
Cu 306.0000 ppm -10.0000 Machine Learning RMSEP Result using a ML process involving SVM, and schemes to select import peaks/spectral featurs
Cu 714.0000 ppm 425.4600 Univariate Calibration Univariate calibration result to demonstrate improvement obtained by ML solution