Quantitative Detection of Chromium Pollution in Biochar Based on Matrix Effect Classification Regression Model

Area of Science:

• Environmental Science
• Analytical Chemistry
• Materials Science

Background:

• Biochar application is a key soil remediation technology in China.
• Accurate heavy metal detection in biochar is crucial to prevent secondary soil pollution.
• Varied biochar compositions present matrix effects challenging accurate elemental analysis.

Purpose of the Study:

• To develop a Laser-Induced Breakdown Spectroscopy (LIBS) method for quantitative chromium (Cr) detection in biochar.
• To address the matrix effect challenge in biochar analysis using a novel classification approach.
• To improve the accuracy and reliability of heavy metal quantification in biochar for environmental safety.

Main Methods:

• Utilized unsupervised hierarchical clustering on elemental LIBS data to classify biochar samples.
• Employed a supervised K-nearest neighbor (KNN) algorithm to assign prediction samples to calibration classes.
• Developed and compared multiple Partial Least Squares Regression (PLSR) models for quantitative Cr analysis.
• Implemented a 3-classification regression model for optimized prediction performance.

Main Results:

• A 3-classification regression model demonstrated superior calibration performance with lower averaged relative standard deviations of cross-validation (ARSDCV).
• The optimized model achieved a low averaged relative standard deviations of prediction (ARSDP) of 8.13%.
• The LIBS approach effectively minimized the impact of complex biochar matrices on Cr quantification.

Conclusions:

• LIBS combined with a matrix effect classification regression model offers accurate Cr quantification in biochar.
• This method provides a reliable tool for assessing biochar quality and ensuring soil remediation safety.
• The developed technique overcomes limitations of traditional methods in analyzing complex biochar matrices.