Scholarlink

<p>Accurate quantification of microplastics (MPs) in soils remains analytically challenging due to complex mineral and organic soil matrices. Microscopy, spectroscopy and thermoanalytical techniques are widely applied to analyse MPs in soils. However, integrated workflows enabling simultaneous assessment of quantitative surface properties remain limited. Surface roughness and complexity may influence particle-environment interactions and are therefore relevant for understanding the environmental behaviour of MPs. This study developed and evaluated an oxidative- and corrosive-substance-free workflow for the simultaneous assessment of MP abundance, size, shape, and quantitative surface roughness and complexity in agricultural soils. The workflow combines density separation with freezing, 3D Laser Scanning Confocal Microscopy (3D LSCM; Keyence VK-X1000, Japan) and machine-learning-based automated detection. In addition to enabling time-efficient particle classification, machine-learning integrates multi-layer 3D LSCM outputs, including height, laser reflection, and colour (RGB) information. Data acquisition was performed at a pixel<sub>xy</sub> size of 2.7&#xa0;μm and a height pitch of 4&#xa0;μm. The method was evaluated using three agricultural topsoils spiked with transparent and black low-density polyethylene and polypropylene fragments (&lt; 53&#xa0;μm, 53–100&#xa0;μm, 100–250&#xa0;μm) and polypropylene fibres (1000&#xa0;μm length). MPs ≥ 53&#xa0;μm were reliably detected with a mean recovery of 80% ± 28% in soils with low to medium particulate organic matter content (POM). As expected, detection performance decreased in soils with high POM content, as POM was not removed during sample preparation, which made it particularly difficult to determine black MPs and fibres. Up to four 25&#xa0;g samples can be processed and analysed within three days, enabling time-efficient MP (≥ 53&#xa0;μm) assessment in agricultural soils and complementing established analytical approaches through quantitative surface characterisation.</p> Graphical abstract <p></p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Microplastics detection in agricultural soil combining 3D Laser Scanning Confocal Microscopy with machine learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年6月17日

查看全文
  • Tabea Scheiterlein,
  • Peter Fiener

摘要

Accurate quantification of microplastics (MPs) in soils remains analytically challenging due to complex mineral and organic soil matrices. Microscopy, spectroscopy and thermoanalytical techniques are widely applied to analyse MPs in soils. However, integrated workflows enabling simultaneous assessment of quantitative surface properties remain limited. Surface roughness and complexity may influence particle-environment interactions and are therefore relevant for understanding the environmental behaviour of MPs. This study developed and evaluated an oxidative- and corrosive-substance-free workflow for the simultaneous assessment of MP abundance, size, shape, and quantitative surface roughness and complexity in agricultural soils. The workflow combines density separation with freezing, 3D Laser Scanning Confocal Microscopy (3D LSCM; Keyence VK-X1000, Japan) and machine-learning-based automated detection. In addition to enabling time-efficient particle classification, machine-learning integrates multi-layer 3D LSCM outputs, including height, laser reflection, and colour (RGB) information. Data acquisition was performed at a pixelxy size of 2.7 μm and a height pitch of 4 μm. The method was evaluated using three agricultural topsoils spiked with transparent and black low-density polyethylene and polypropylene fragments (< 53 μm, 53–100 μm, 100–250 μm) and polypropylene fibres (1000 μm length). MPs ≥ 53 μm were reliably detected with a mean recovery of 80% ± 28% in soils with low to medium particulate organic matter content (POM). As expected, detection performance decreased in soils with high POM content, as POM was not removed during sample preparation, which made it particularly difficult to determine black MPs and fibres. Up to four 25 g samples can be processed and analysed within three days, enabling time-efficient MP (≥ 53 μm) assessment in agricultural soils and complementing established analytical approaches through quantitative surface characterisation.

Graphical abstract