Laser-induced breakdown spectroscopy (LIBS), as an elemental composition analysis technique, has many unique advantages and great potential for applications in water detection. However, the quality of LIBS spectral signals, such as signal-to-noise ratio and stability, is often poor due to the matrix effects of water, limiting its practical performance. To effectively remove the inherent weak radiation in experimental spectral data that can be easily mistaken for noise, this paper proposes a denoising algorithm for processing spectral data using a self-built blank sample spectral database of deionized water samples, and designs a complete data processing workflow. It includes steps such as blank sample data screening, internal standard correction, blank sample correction, and spectral smoothing. Against the backdrop of marine applications, experimental spectral data for target elements Na, Mg, Ca, K, Sr, and Li were processed with this algorithm. The results show that after algorithm processing, the spectral quality was significantly improved, with the signal-to-noise ratio and detection limits of various elements improved by at least one order of magnitude. The signal-for Li increased by up to 36 times, and the detection limit for K decreased by up to 25.2 times. Additionally, tiny spectral peaks that could not be observable in the original spectral data could be effectively extracted after processing. From a technical implementation perspective, the database establishment and data process are simple and practical, with universal applicability. Therefore, this method has good potential and wide foregrounds in many other water sample LIBS detection technologies.
Keywords: Blank sample; Data processing; LIBS; Noise reduction; Water sample.
Copyright © 2024 Elsevier B.V. All rights reserved.