Research on rapid and accurate identification of steel grades based on laser-induced breakdown spectroscopy and restricted Boltzmann machine-back propagation algorithm

Ginseng, a traditional Chinese herbal medicine, has experienced increasing application. Due to variations in the medicinal value of ginseng based upon its cultivation location, an accurate method is required to identify its origin. Here laser-induced breakdown spectroscopy (LIBS) and Raman spectroscopy were employed with machine learning algorithms for the authentication of ginseng. Accurate origin identification was conducted on dried white ginseng. Initially, LIBS and Raman spectra were acquired separately. Regional differences were not observable in the LIBS spectra. Conversely, in the Raman spectra, similarity was observed between the Baishan and Dunhua regions, contributing to reduced accuracy. Subsequently, the LIBS and Raman spectra were fused. Principal component analysis (PCA) was employed to reduce the dimensionality of the LIBS, Raman, and fused data. The top 20 principal components for these data sets explained variances of 97.17%, 86.51%, and 94.22%, respectively. The reduced-dimension LIBS, Raman, and LIBS-Raman data were utilized as input variables for random forest, extreme gradient boosting, and categorical boosting (CatBoost) classification. The results indicate that the LIBS-Raman data yield the highest accuracy with CatBoost reaching 99%. Compared to the results using LIBS and Raman, there were improvements in accuracy of 11% and 6%. This method holds significance for the rapid and accurate identification of ginseng.

Rapid detection and identification of objects using a self-designed methodology based on LIBS and PCA-DVSM – taking rosewood for example

This study presents the application of laser-induced breakdown spectroscopy (LIBS) for analyzing various copper-bearing critical ores with significant Cu concentrations. LIBS detected Cu as a base element, along with other minor elements including Al, C, Fe, Mg, Ni, Si, and Zn, under optimized experimental conditions that include 80±0.3mJ laser energy, 2 μs delay time, ∼500μm spot size, and a 45° angle between the collecting lens and the sample surface. The energy-dispersive X-ray technique was employed to determine the elemental concentrations and spatial distributions within the sample, based on Kα, Kβ, and Lα characteristic lines. Quantitative analysis in LIBS is challenging due to matrix effects on line intensities, often requiring matrix-matched standards; however, the multielemental quality of LIBS spectra enables the detection of matrix types for accurate classification. In this contribution, we applied an unsupervised principal component analysis (PCA) on pre-processed LIBS data to reduce dimensionality and visualize clusters, showing that the first three principal components (PCs) account for 97.9% of the total variance (PC1: 69.8%, PC2: 20.3%, PC3: 7.8%). Elliptical PCA clustering with a 96% confidence interval was achieved using SIMCA. A supervised partial least squares-discrimination analysis model is used to identify the variables that contribute most to classification. The model yields cumulative X and Y variances of 97.86% and 99.96%, respectively, with an R2 range of 0.83-0.99 across the first 6 factors. Furthermore, LIBS 2D mapping is carried out using Cu spectral lines at 510.6 (2P3/2→2D5/2), 515.3 (2D3/2→2P1/2), and 521.8nm (2D5/2→2P3/2), and Zn at 481.1nm (3S1 → 3P2), over 50 and 200 scans to visualize the element spatial distribution. Mapping is cross-validated using Pearson's correlation covering a 50×50mm2 area, achieving ∼150μm spatial resolution and an average root mean PRESS of ∼94% with a high correlation of ∼0.989. The results show the efficiency of LIBS integrated with multivariate methods for pattern recognition, classification, and spatial analysis in the exploration of copper ores.

In this study, classification models based on laser-induced breakdown spectroscopy (LIBS) technology, combined with machine learning (ML) and deep learning (DL) algorithms, were proposed to enable the rapid and high-precision classification of small specimen uranium ores. LIBS spectral data from 12 types of uranium ore samples were collected and subsequently subjected to standard normal variate (SNV) preprocessing before model construction. The classification models were constructed using the random forest (RF) algorithm and three DL algorithms-feedforward neural network (FNN), convolutional neural network (CNN), and long short-term memory (LSTM)-incorporating two feature extraction methods: least absolute shrinkage and selection operator (LASSO) and principal component analysis (PCA). The classification performance and generalization ability of the different models and feature extraction strategies were systematically evaluated. It was found that the RF model exhibited significant overfitting when the training set size was small, and its performance improvement required an increase in training set proportion. When LASSO feature selection was incorporated, the DL models outperformed the RF model, although some misclassifications were still observed. In contrast, PCA, which utilized only the first five principal components, effectively retained the global discriminative information of the spectra. All DL models based on PCA features achieved 100% classification accuracy for both the training and testing sets. This study demonstrates that PCA can effectively extract global spectral information, overcoming the limitations posed by small specimens in LIBS classification tasks for uranium ores. When combined with DL algorithms, PCA significantly improves classification performance and generalization ability, offering a reliable technical pathway for the rapid and accurate identification of small specimen uranium ores.

Laser-induced breakdown spectroscopy and multivariate statistics for the rapid identification of oxide inclusions in steel products

Feature-enhanced dual-input transformer for LIBS quantitative analysis of minor-content elements.

Análisis de plaguicidas en hortalizas de hojas verdes mediante espectroscopía de plasmas producidos por láser

The Manuguru geothermal area, located in the Telangana state, is one of the least explored geothermal fields in India. In this study, characterization of the soil samples is carried out by laser-induced breakdown spectroscopy (LIBS) coupled with analytical spectral-dependent principal component analysis. A total of 20 soil samples were collected both from near the thermal discharges as well as away from the thermal manifestations. LIBS spectra were recorded for all the collected soil samples and principal component analysis (PCA) was applied to easily identify the emission lines majorly responsible for variety classification of the soil samples. In this submission, a modified PCA was developed which is based on the spectral truncation method to reduce the huge number of spectral data obtained from LIBS. The PCA bi-plot on the LIBS data reveals the presence of two different clusters. One cluster represents the soil samples collected from the close vicinity of the thermal manifestations whereas the other cluster contains the soil samples collected away from the thermal sprouts. PCA performed on the chemical dataset of the soil samples also reveals the same clustering of the soil samples. Both LIBS and chemical analysis data shows that soil samples near the thermal waters are found to be enriched in B, Sr, Cs, Rb, Fe, Co, Al, Si, Ti, Ru, Mn, Mg, Cu, and Eu concentrations compared to the soil samples located away from thermal manifestations. This study demonstrates the potential use of LIBS coupled with PCA as a tool for variety discrimination of soil samples in a geothermal area. LIBS is shown to be a viable real-time elemental characterization technology for these samples, avoiding the rigorous dissolution required by other analytical techniques.

Laser Induced Breakdown Spectroscopy (LIBS) is an analytical technique, used to classify and potentially quantify elements in complex hosts (or matrices). Vacuum Ultraviolet Laser Induced breakdown Spectroscopy (VUV LIBS) can offer potential improvements in detection of light elements in bulk metals over traditional LIBS in the visible region. This is due to presence of an abundance of resonance transitions at shorter wavelengths. This extends the ability to discriminate between the emission from different elements, particularly light elements such as carbon, sulfur, lithium, beryllium etc. Additionally, the precision of LIBS is limited by the continuum emission at the early stage of the plasma lifetime. The performance of LIBS can be improved by using a time resolved detection system [5], reducing the contribution from the continuum. In this study, the detection of the carbon content in steel samples is performed by time- integrated and time-resolved VUV LIBS. The experimental setup consists of a dual pulse system with Nd:YAG laser (1064 nm, up to 450 mJ, pulse duration 6 ns) used to irradiate the samples, a vacuum system to prevent absorption of the VUV radiations and a VUV spectrometer to collect and measure the emission spectra. Samples of four different concentrations of carbon in steel are used for the study. The resultant time integrated LIBS limit of detection and signal to background ratio is compared with time resolved VUV measurements.

The rapid detection and identification of the electronic waste (e-waste) containing rare earth (RE) elements is of great significance for the recycling of RE elements. However, the analysis of these materials is extremely challenging due to extreme similarities in appearance or chemical composition. In this research, a new system based on laser induced breakdown spectroscopy (LIBS) and machine learning algorithms is developed for identifying and classifying e-waste of rare-earth phosphors (REPs). Three different kinds of phosphors are selected and the spectra is monitored using this new developed system. The analysis of phosphor spectra shows that there are Gd, Yd, and Y RE element spectra in the phosphor. The results also verify that LIBS could be used to detect RE elements. An unsupervised learning method, principal component analysis (PCA), is used to distinguish the three phosphors and training data set is stored for further identification. Additionally, a supervised learning method, backpropagation artificial neural network (BP-ANN) algorithm is used to establish a neural network model to identify phosphors. The result show that the final phosphor recognition rate reaches 99.9%. The innovative system based on LIBS and machine learning (ML) has the potential to improve rapid in situ detection of RE elements for the classification of e-waste.

This study focuses on the discrimination of extracted animal fats in liquid form using laser induced breakdown spectroscopy (LIBS) technique assisted with principal component analysis (PCA). The interaction of laser and liquid sample produces liquid splashing due to strong shock wave effect and subsequently generates lower intensity of LIBS signals. LIBS difficulties in liquid are resolved using paper substrate to enhance LIBS emission intensity. Laser pulse from Q-switched Nd:YAG laser with energy of 220 mJ and frequency of 1 Hz was used to ablate extracted animal fats. The obtained LIBS spectra of extracted animal fats were further evaluated using PCA. LIBS spectra are compressed and visualised as data points in the score plot of PCA. PCA results demonstrated that data points from different extracted animal fats were clustered separately in the score plot with variance greater than 90%. The findings show LIBS system assisted with PCA was capable to differentiate various extracted animal fats.

The identification of plastics is of great importance to recycling companies, as the littering of plastic wastes has rapidly increased because of the extensive use of plastics. To maintain the economics of recycling extremely large volumes of waste materials, rapid and accurate identification of these plastics is crucial. In this study, we demonstrate the efficacy of laser‐induced breakdown spectroscopy (LIBS) in the rapid identification of high‐density polyethylene (HDPE) and low‐density polyethylene (LDPE) used for toy manufacturing. For data analysis of the LIBS spectra, multivariate data analysis using principal component analysis (PCA) was utilized. The analyses of the data clearly showed that the elemental and molecular information obtained from LIBS is effective in the identification of three types of polyethylene. The proposed method was successfully applied and can be used as an alternative to the traditional methods used for analyses of plastics.

Combination of laser-induced breakdown spectroscopy and Raman spectroscopy for multivariate classification of bacteria

HyperPCA: A powerful tool to extract elemental maps from noisy data obtained in LIBS mapping of materials

Laser-induced breakdown spectroscopy (LIBS), as a localized remote sensing technique, is an emerging method for analyzing the composition of soil samples. However, studies leveraging LIBS spectral data to explore soil heterogeneity across different depths remain limited. In this study, we applied LIBS combined with machine learning to estimate the presence of some chemical elements at varying depths. The analysis detected significant concentrations of elements such as Ca, Mg, and N, revealing spatial heterogeneity in nitrogen content across ten soil layers. Back propagation (BP) neural network, random forest algorithm, and convolutional neural network were applied for soil depth classification and regression analysis. BP achieved the highest F1 score of 0.9125 in classification and the lowest root mean squared error (RMSE) of 9.8663 cm in regression, demonstrating its superiority in soil spectral analysis. Dimensionality reduction was performed using principal component analysis, linear discriminant analysis (LDA), and SelectKBest, followed by regression with BP. LDA combined with BP achieved the best performance, with an RMSE of 8.7742 cm. The integration of LIBS, LDA, and BP provides an efficient solution to contribute to a rapid and precise identification of different abundant chemical elements in the soil at different depths.