Sample pre-treatment
Drinking water
Pre-treatment is adapted based on the selected detection limit. For larger particles >10 µm, water samples are filtered through 0.2 µm aluminium oxide filter membranes to concentrate particles for consequent analysis. For samples requiring additional preparation, hydrogen peroxide (H₂O₂) digestion is applied to remove organic residues, and hydrochloric acid (HCl) digestion may be used to dissolve carbonate content. The filtered residues are then analyzed directly by automated Raman microspectroscopy.
 
Contaminated water and sludge
Pre-treatment of contaminated water involves filtering the sample to recover suspended solids. Organic matter is digested using hydrogen peroxide (H₂O₂) as the primary oxidizing agent. For samples with higher organic content, sodium hypochlorite (NaClO) or Fenton’s reagent (a combination of H₂O₂ and Fe²⁺ catalyst) may also be used. To separate microplastics from denser particles, density separation with zinc chloride (ZnCl₂) is performed. The remaining residue is filtered and prepared for microspectroscopic analysis.
 
Spectral acquisition
Samples are analyzed by Raman microspectroscopy within a representative subsample area, defined by the spatial limit of detection (particle size cut-off). Each sample undergoes automated Raman microspectroscopy for 12–24 hours, generating spectral and morphological data for tens of thousands of individual particles. Raman measurements are conducted at 20°C using a Horiba LabRAM Soleil (Jobin Yvon, France). Samples are excited at 8% (7.2 mW) power output with a high-stability, air-cooled He–Cd 532 nm laser diode and a Nikon LV-NUd5 100× objective. The unpolarized confocal laser beam provides a lateral resolution of approximately 1 µm. Spectra are collected in the range of 200–3400 cm⁻¹ using a 600 grooves/cm grating with a 100 µm slit, achieving a spectral resolution of approximately 1 cm⁻¹. Particle analysis within each mosaic, constructed with the LabSpec6 (LS6) SmartView configuration, is performed using the Particle Finder application V2. LS6 SmartView determines the topography (±50 µm), saves the focal points of all particles in the micrograph, and enables rapid refocusing of the stage on the relevant particles.
 
Spectral matching and verification
Using Spectragryph spectral analysis software V1.2.17d (Dr. Friedrich Menges SoftwareEntwicklung, www.effemm2.de/spectragryph), raw spectra are processed with adaptive baseline correction at 15% coarseness. The processed spectra are cross-referenced across their entire spectral range using an in-house library containing selected spectra from the SLoPP and SLoPP-E libraries (Munno et al., 2020) and the Cabernard library (Cabernard et al., 2018), along with self-obtained in-house polymer spectra. Spectral matches are evaluated using hit quality index (HQI) values ranging from 0 to 100%. Spectra with HQI values above 65% are classified as microplastic candidates and are manually inspected and validated by a trained interpreter.
 
Correction of false positives (unintentionally partitioned particles)
Due to the frame-by-frame acquisition process required for obtaining accurate Raman measurements of microplastics as small as 1 µm in diameter, particles located on the edges of a frame (edge-particles) are often partitioned and misidentified as multiple particles. At the current microscopic resolution, each frame measures 60 × 40 µm, resulting in an overestimation of smaller size fractions and an underestimation of larger particles. While the LS6 Particle Finder application includes an option to exclude edge particles, this approach disproportionately removes larger particles, which are more likely to intersect with the frame's edge, leading to underestimation of the true MP concentration. To address this, a custom script (Microplastic Solution, France) was developed to improve data post-processing by merging unintentionally partitioned microplastics. Although the code cannot be shared due to commercial interests, its principal functions are disclosed. The script calculates the distance between particles of the same polymer type to identify overlaps, defined as the distance between particle centers being smaller than the major axis of one of the particles. Overlapping particles are grouped, and the particle with the highest spectral hit quality index (HQI) is designated as the "leader." The leader's index defines the new, merged particle, which inherits the cumulative area of the group, while the remaining particles are discarded. The diameter of the merged particle is calculated as the area-equivalent diameter, assuming a circular shape.
 
Negative- and positive quality controls
Blank control and correction
To ensure reliable results, a procedural blank experiment is conducted to estimate and correct for contamination introduced during sample pre-treatment. A contamination-free solution is prepared and treated according to the same protocol as the actual samples. During analysis, particles identified in the procedural blank are carefully characterized by polymer type and size. Blank correction is performed by comparing microplastics in the primary sample with those in the blank. Particles are matched ("paired") if they share the same polymer type and their sizes fall within a defined range. Once paired, the corresponding particle in the primary sample is subtracted, and the matching blank particle is excluded from further corrections. This process ensures that procedural contamination is accurately accounted for without overestimating its impact. Blank correction is applied uniformly across all samples, with equivalent portions of the filter area analyzed to maintain consistency. This approach allows contamination to be effectively identified and subtracted, ensuring the validity of the results. We provide both blank corrected and non-blank corrected data.
 
Recovery control and correction
To account for unintentional microplastic loss during sample pre-treatment, a procedural recovery experiment is conducted using EasyMP™ red polyethylene (PE) fragments (Microplastic Solution, France). Known quantities of these environmentally relevant fragments are added to test samples, which are then processed using the same protocol as the actual samples. After processing, the remaining number of spiked microplastics is evaluated to determine the analytical recovery rate.
EasyMP™ fragments allow for the assessment of recovery across different size ranges, providing insight into trends such as lower recovery for larger particles and, in some cases, higher-than-expected recovery for smaller particles due to potential fragmentation during pre-treatment. Based on the findings, a mathematical model is developed to describe the relationship between particle size and recovery rate. This model is then applied to correct the analytical recovery of all detected microplastics within size groups, ensuring accurate quantification of microplastics in the processed samples.