基于算法优化的便携式X射线荧光光谱仪土壤重金属定量检测模型的建立

将小波变换去噪和迭代多项式拟合去除本底结合,对便携式X射线荧光光谱仪(PXRF)谱图进行去除噪声和本底扣除处理,然后根据土壤重金属含量标准值和处理后测定的计数建立各重金属的标准曲线。相比较未做去噪和本底扣除,经过sym4小波去噪和9次迭代多项式拟合处理后的仪器标准曲线决定系数R²范围提升至0.965 2~0.998 5,有效提高了仪器检测的正确度和精密度。采用算法优化(小波变换去噪+迭代多项式拟合)前后的便携式X射线荧光光谱仪分别对土壤样品中的铜、砷、铬、锌、铅和镍等6种重金属元素含量进行测定,分别从检出限、正确度和精密度等方面进行了对比分析。结果表明:在正确度方面,未处理前的相对误差范围为0.2%~4.8%,而仪器算法处理后测定样品的相对误差范围为0.4%~4.6%;在精密度方面,仪器算法处理后相对标准偏差范围为0.50%~5.2%,未处理前相对标准偏差范围为0.80%~15%;在检出限方面,仪器算法处理后分析6种重金属检出限范围低至3.5~22.0 μg/g。最后通过与实验室检测结果对比,数据显示部分元素均能接近实验室分析标准,实验结果无显著性差异,具有可比性,可用于大面积土壤环境修复和环境监测。

Establishment of quantitative detection model for heavy metals insoil using portable X-ray fluorescence spectrometerbased on algorithm optimization

Combining wavelet transform denoising with iterative polynomial fitting to remove background, the spectra of portable X-ray fluorescence (PXRF) were treated with noise removal and background deduction. Then the standard curve of each heavy metal was established according to the standard value of soil heavy metal content and the counting rate measured after treatment. Compared to the method without noise removal and background deduction, the determination coefficient (R²) range of the standard curve after sym4 wavelet denoising and 9 times iterative polynomial fitting were improved to 0.965 2-0.998 5, which effectively improved the accuracy and precision of the instrument detection. Moreover, the contents of six heavy metal elements (including copper, arsenic, chromium, zinc, lead and nickel) in soil samples were determined using PXRF with and without algorithm optimization (wavelet transform denoising iterative polynomial fitting). The detection limit, accuracy and precision were compared and analyzed. The results showed that: in terms of accuracy, the relative error range before algorithm optimization were 0.2%-4.8%, and it was improved to 0.4%-4.6% after algorithm optimization treatment; in terms of precision, the relative standard deviation range was 0.50%-5.2% after algorithm optimization, while it was 0.80%-15% before algorithm optimization; in terms of detection limit, the detection limits of 6 heavy metals were reduced to 3.5-22.0 μg/g after algorithm processing. Finally, compared with the laboratory test results, the data showed that the results of some elements could be close to the laboratory analysis standard, and the experimental results have no significant difference. The testing results were comparable. The proposed method could be used for large area soil environmental remediation and environmental monitoring.