Towards quality-assured measurements of microplastics in soils using fluorescence microscopy

Fluorescence microscopy utilising Nile Red staining was applied and tested to analyse microplastic particles in soil. Given a large number of analytical methods for measuring plastic particles in environmental and biological matrices, often utilising microscopy/spectroscopy approaches, efforts were made to validate this method for a range of soils and different microplastic (MP) particle types (varying size and morphology).  Eight MP types (including both non-biodegradable and biodegradable plastics) within three size categories (dia. 500-1000 µm, 100-250 µm and 10-150 µm) were spiked into three different agricultural soils (loam-reference soil, clay, and sandy soils) for a comprehensive assessment of the fluorescence microscopy methodology. Each soil (with replicates) was subject to digestion, density separation and filtration (with Nile Red staining) prior to analysis using fluorescence microscopy (GFP filter set, excitation/emission 470/525 nm) and bright filed microscopy for black microplastics. As a tool to shorten the analysis time, a digital image analysis pipeline using Image J was developed, allowing fully automated particle recognition and quantification of MPs in the samples. The main steps for image analysis, including background correction, fluorescent intensity thresholding, watershed segmentation and particle analysis, were optimised, the validation of which showed high accuracy (88% match to true observation) for MP particles on a filter without a soil matrix. To avoid false positive results due to the presence of natural organic particles from the soil matrix, the numbers of microplastics recovered from spiked samples were corrected with those found in the non-spiked soils. Recoveries ranged from 80-90% for MP with sizes from 500-1000 µm regardless of the soil types, whereas those for smaller MP (10-250 µm) varied between different soils and plastic types (e.g., recovery for low-density polyethylene, LDPE, from sand and loam-reference soil were 85% and 90% respectively whereas those for polybutylene adipate terephthalate/polylactic acid, PBAT/PLA, were 60% and 10% respectively). The lowest recovery rates were observed in clayey soil (20% for LDPE and 5% for biodegradable plastics PBAT/PLA). A relationship between the microplastic mass (LDPE and PBAT/PLA fragments) and the corresponding particle number was established in this study, which enabled the conversion between mass and particle number data. Fluorescence microscopy with Nile Red staining and automatic particle recognition software provides a relatively reproducible and accurate technique for plastic particle counts in heterogeneous matrices like soil. Still, selectivity for different polymer types is clearly limited.