Agricultural soil survey and inter-laboratory comparison of microplastics across China, Egypt, India, Sri Lanka, and Vietnam, 2022-2025

Part 1 - Soil Survey:
- pH and electrical conductivity: Soil samples (n=1) were analysed in a 1:2.5 soil:distilled water solution using standard electrodes. Data recorded in paper notebook, then transferred to Excel file. Excel file exported as csv file.

- Nutrients (ammonium, nitrate, DOC, phosphate): Soil samples (n=1) were analysed in a 1:5 soil:0.5 M K2SO4 solution (ammonium, nitrate, DOC) using colorimetric analysis for ammonium and nitrate, and a TOC analyser for DOC. Soil samples were analysed in a 1:5 soil:0.5M acetic acid or 0.5M sodium bicarbonate solution (phosphate) using colorimetric analysis. Data was exported as Excel files from respective instruments, then exported as csv file.
- Microplastic quantification: Microplastics were extracted from soil samples in each country following a standardised protocol[2]. Briefly, soil samples were dried at 40 °C for min. 24 h, larger aggregates were broken up, and plant material was removed before weighing out 5 ± 0.05 g of dry soil into a clean glass beaker. For density separation, 230 ml saturated NaCl (density 1.2 g cm-3) were added to the beaker, the solution was homogenised for 30 min using a magnetic stirring plate, and left to stand overnight. Following this, the supernatant was decanted into a new glass beaker, and 4 M NaOH was added to a fixed volume of 500 ml, after which the solution was homogenised and left to stand overnight to facilitate digestion of biologically-derived organic material. The supernatant was then decanted into a new glass beaker and 250 µl of Nile Red solution added, homogenised, and incubated in the dark at room temperature for 30 min before vacuum filtration onto a suitable membrane. Membranes were stored in glass petri dishes covered in aluminium foil and analysed within 24 h. Particle quantification on membranes was carried out using fluorescent microscopy within each country. Between 5 - 100 % of the membrane area were analysed, depending on country, and resulting images were processed using ImageJ (v. 1.54f). Briefly, a Macro was customised for each country, scaling images accordingly, applying a suitable colour threshold with the default thresholding method, adjusting the brightness level to a suitable setting (according to the average best fit of 10 randomly selected images), and exporting particles' Feret's diameter. Data was filtered to only include particles with Feret's diameter of 20 - 500 µm, and scaled to an equivalent particle number of 100% membrane area (= 5 g dry soil). Total particle number was then divided by 5 to express particles g-1 dry soil.

- Microplastic identification: Complimentary to the particle quantification, fluorescent particles (i.e. suspected MPs) were also analysed using IR spectroscopy within each country. Particle spectra were matched against instrument-specific or open-source libraries with an average match quality for polymers of 40.6% (China), 80.8% (Egypt), 70.6% (India), 85.7% (Vietnam), and 84.7% (Sri Lanka). China, Egypt, India, and Vietnam manually selected individual particles for analysis, resulting in an average of 10 particles analysed across these countries due to resource constraints and instrument availability. Data for China are only available for 0 - 10 cm depth. Sri Lankan samples were analysed in the UK, utilising an automated analysis of particles per subsample (analysing approx. 1% of total sample volume). A match quality threshold of ≥ 80% was applied to the Sri Lankan dataset only, according to manufacturer's guidance. Due to no clear manufacturers' instructions and the already low manual particle count per sample in the other countries, no match quality threshold was applied.

- Environmental data: Precipitation and air temperature data was sourced from the nearest weather station for each location.

- Farm metadata: Data regarding farm practices was collected directly from participating farmers, written on paper, entered into an Excel file, exported as a csv file.
Part 2 - Inter-laboratory standard comparison
- Microplastic quantification: The standards consisted of three matrices: silicon dioxide (SD; quartz sand), bentonite clay (BT), and country-representative soil; all with and without 0.04% (w/w) polyethylene (PE; 34-50 μm particle size). SD and BT were selected to represent two commonly found soil textures (sand, clay), and soil was selected representative of each country's prevalent soil types. The soil was collected by each laboratory and spiked with PE (provided by UK) to 0.04% (w/w). The UK also determined the number of PE particles in the mass required to achieve the 0.04% (w/w) concentration in 5 g matrix to estimate the average particle number per sample and therefore estimate maximum particle number recoverable. All standards were analysed in triplicates alongside blanks (no matrix, only reagents), following a standardised extraction and analysis protocol [2]. The same methodology was applied as described above.
- Microplastic identification: The standards consisted of 9 individual polymer types (PET, PP, PTFE, PU, PC, PE, PS, PA, PVC) with different size ranges (Table 9). Each country identified 10 randomly selected particles per standard. The analysis was carried out blind, i.e. participating countries (apart from UK) were unaware of what polymer types were included in each standard vial. Sri Lanka had no capability to analyse polymer standards via IR spectroscopy and was therefore excluded from this analysis. India was unable to analyse 'standard 3' (PTFE) due to sample loss. In addition to individual standards, each laboratory received a homogenised polymer mixture containing the same nine polymers in known weight proportions (Table 10). Each laboratory analysed three replicates of the mixed polymer standard, selecting 50 random particles per replicate. India was unable to analyse Mix 1 and 2 due to sample loss, therefore only Mix 3 is included.