Quantitative Analysis of Elements in Fertilizer Using Laser-Induced Breakdown Spectroscopy Coupled with Support...

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Chemometrics

Background:

• Rapid elemental analysis is crucial for controlling compound fertilizer production.
• Accurate detection of nitrogen (N), phosphorus (P), and potassium (K) is vital for the fertilizer industry.

Purpose of the Study:

• To compare the detection capabilities of laser-induced breakdown spectroscopy (LIBS) coupled with support vector regression (SVR).
• To establish an accurate and reliable method for the rapid quantitative determination of N, P, and K in compound fertilizers.

Main Methods:

• Utilized LIBS for elemental analysis of 58 fertilizer samples.
• Employed SVR with four parameter optimization methods: grid search (GSM), genetic algorithm (GA), particle swarm optimization (PSO), and least squares (LS).
• Applied the Kennard-Stone (KS) method for sample set division.

Main Results:

• Evaluated models based on training time, determination coefficient, and root-mean-square error.
• The LIBS technique combined with least squares-support vector regression (LS-SVR) demonstrated high accuracy and reliability.
• LS-SVR proved effective for quantitative determination of N, P, and K in complex fertilizer matrices.

Conclusions:

• LIBS coupled with LS-SVR is a robust method for rapid elemental analysis in fertilizers.
• This approach provides a reliable tool for quality control in the compound fertilizer industry.
• The study validates the effectiveness of chemometric methods for complex sample analysis.