Back to Search
Start Over
Collaborative estimation of heavy metal stress in wheat seedlings based on LIBS-Raman spectroscopy coupled with machine learning.
- Source :
- JAAS (Journal of Analytical Atomic Spectrometry); Oct2023, Vol. 38 Issue 10, p2059-2072, 14p
- Publication Year :
- 2023
-
Abstract
- Heavy metal lead (Pb) accumulation in agricultural soil has become a serious threat to human health through the food chain in recent decades. Wheat, being one of the most important crops, is a crucial source of food for humans. Pb in soil can pose a significant risk to the wheat grain, as it can be transported through the root and stem. This transportation not only increases the Pb concentration in wheat, but also threatens the wheat's overall food security. Therefore, it is imperative to investigate the effects of Pb stress on wheat seedlings. Double-pulse laser-induced breakdown spectroscopy (DP-LIBS) and surface-enhanced Raman spectroscopy (SERS) as promising enhanced spectroscopy techniques were coupled to estimate Pb stress in wheat seedlings. Wheat seedling samples under Pb stress at 6 concentration gradients (0(c.k.), 500, 1000, 2000, 3000, and 5000 mg L<superscript>−1</superscript>) were cultivated with the same substrate and management. The SERS and DP-LIBS spectra of the root, stem and leaf were obtained by using spectrometers in an enhanced way. Principal component analysis (PCA) was employed for preliminary analysis of DP-LIBS and SERS data to investigate the clustering of samples, including substrates and the wheat seedling root and leaf, under different Pb concentration stresses. However, while PCA enabled the rough differentiation of the samples, the classification accuracy requires improvement. The pre-processed SERS, DP-LIBS, and fusion spectral data (LIBS-Raman) were analyzed using discriminant models such as least squares support vector machines (LS-SVM), multi-layer perceptron-artificial neural network (MLP-ANN), radial basis function-artificial neural network (RBF-ANN) and probabilistic neural networks (PNN). The accuracy of the model built by LIBS-Raman was significantly better than the model built based on SERS and DP-LIBS data alone. LIBS-Raman increased the accuracy of the discrimination model, which highlights the effectiveness of collaborative spectroscopic methods for heavy metal detection. This study investigated the migration of Pb in the substrate-root-leaf. At 0 concentration (c.k.), Pb was not detected in the substrate and wheat seedling root and leaf. At the same concentration, the level of Pb stress in the root was higher than that in the leaf. This might be attributed to the self-protection mechanism of wheat seedlings, which had a weak ability to transport Pb ions. This work provides theoretical support for collaborative spectral data, furnishes a reference for the development of a combined spectrometer, and offers a valuable method for monitoring heavy metal contamination in plants. [ABSTRACT FROM AUTHOR]
Details
- Language :
- English
- ISSN :
- 02679477
- Volume :
- 38
- Issue :
- 10
- Database :
- Complementary Index
- Journal :
- JAAS (Journal of Analytical Atomic Spectrometry)
- Publication Type :
- Academic Journal
- Accession number :
- 172784106
- Full Text :
- https://doi.org/10.1039/d3ja00243h