The scale of the problem the method addresses is enormous. Global plastic production has climbed from roughly two million metric tons in 1950 to nearly 391 million metric tons in 2021, and because most plastics do not biodegrade, discarded items persist and fragment. Macroplastics larger than 25 millimeters break down under sunlight and physical weathering into mesoplastics and then microplastics, defined as solid polymer-containing particles between one micrometer and five millimeters. These particles are classified as primary, when manufactured at microscopic sizes, or secondary, when produced by the fragmentation of larger debris, and they display enormous diversity in polymer type, shape, color and size.

Aquatic systems bear the brunt of this pollution. An estimated 4.8 to 12.7 million metric tons of plastic waste entered the oceans from land-based sources in 2015 alone, and rivers act both as sinks where particles settle and as major highways transporting them to marine environments. Once in the water, microplastics harm organisms that ingest them, disrupting feeding, energy metabolism and reproduction, and they can act as vectors for persistent organic pollutants and heavy metals that bioaccumulate up the food chain. Understanding where and when microplastics accumulate in river sediments is therefore essential for environmental risk assessment, yet no standardized, robust protocol for sampling and analysis in complex matrices has existed, meaning field studies often produce results that cannot be compared.

The difficulty lies in the sediment itself. Riverbed material is a mixture of light organic matter, such as plant detritus and locally produced biomass, and fine inorganic particles like silt and clay, all of which can trap microplastics and interfere with their chemical identification. Traditional protocols rely on sieving, density separation and filtration, often supplemented by chemical oxidation or enzymatic digestion. Some multi-step procedures applied to organic-rich sediments have required as long as 24 days per batch. Faster alternatives exist, including dye staining with Nile Red, electrostatic or magnetic separation, and oil-assisted extraction, but each carries drawbacks: staining produces false positives in organic-rich samples and adds fluorescence that overwhelms Raman signals, while olive oil residues left on filters proved fluorescent and blocked spectroscopic identification in the Florence team’s own tests.

To overcome these limitations, the researchers combined several established techniques into a streamlined workflow tailored for Raman analysis. Sediment collected from the Arno River in Florence and Pisa was dried, sieved, and sonicated to expose embedded particles. Fenton oxidation, using hydrogen peroxide with an iron catalyst at 35 degrees Celsius, then digested the organic matter over about a week. A density separation in saturated sodium chloride solution floated the low-density fraction, and finally a custom-made glass separator performed an affinity extraction in which microplastics preferentially migrate into cyclohexane, a non-polar solvent, while residual vegetable matter stays in the water phase. A small amount of sodium dodecyl sulfate was added to weaken the adhesion between particles and the sediment matrix, and the extraction was repeated three times to maximize recovery.

The choice of cyclohexane was the key innovation. Unlike olive oil, it does not stick to glassware, evaporates easily, and is optically transparent in the visible range because it is a saturated, symmetrical hydrocarbon with no conjugated electrons or heteroatoms. That means it generates no fluorescence background in Raman measurements, preserving the sensitivity needed to identify even weakly scattering environmental particles. It is also chemically inert, so free radicals left over from the Fenton step are unlikely to degrade it into fluorescing residues. Although cyclohexane is ranked as problematic from a safety and environmental standpoint, the protocol uses only about six milliliters per sample, and the custom glassware was designed to minimize solvent waste.

Validation was carried out on real Arno sediments spanning a wide range of conditions, from fine and medium silts to coarse gravel, with organic carbon contents between 0.14 and 2.2 percent. Samples were spiked with reference particles of five polymers covering a range of densities: polyethylene, acrylonitrile butadiene styrene, polyvinyl chloride, polyamide and polyethylene terephthalate. The overall recovery was 0.65, rising to 0.82 for low-density polyethylene and 0.72 for medium-density PVC, but dropping to 0.42 for high-density PET, which barely floats in the 1.2 grams per cubic centimeter sodium chloride solution. The team accepted this trade-off deliberately, since heavier salts such as sodium iodide or zinc chloride are costly, toxic or energy-intensive to recycle, and the vast majority of environmental microplastics have densities below the threshold. Only about 13 percent of global plastic production exceeds it.

The tests also revealed how strongly the sediment matrix itself influences recovery. Recovery was highest in medium and coarse sands but fell sharply in fine sands and coarse silts, where the greater surface area, cohesiveness and aggregation of small particles entrap microplastics and prevent them from detaching and floating. The cyclohexane extraction step alone achieved an overall efficiency of 0.89, and recovery correlated with polymer wettability: particles with a water contact angle above roughly 75 to 80 degrees were recovered almost completely, while more hydrophilic surfaces were harder to extract. Fibers proved slightly harder to recover than spheres or fragments. Crucially, a stability test confirmed that the Fenton oxidation did not alter the Raman spectra, dimensions or morphology of the reference particles.

When applied to unspiked Arno River samples, the workflow identified five polymers: polypropylene, polyester, polyethylene terephthalate, polycarbonate and polyethylene. Concentrations ranged from 0.030 items per gram of dry sediment in fine sand to 0.653 items per gram in coarse gravel, with a median of 0.151 and a mean of 0.217 items per gram. These values are lower than figures previously reported for emerged sediment bars on the Arno, which trap particles during floods and were measured with the hot needle method known to overestimate counts through false positives, but they align well with studies of smaller Tuscan rivers. Blank samples contained no microplastics above 20 micrometers, so no contamination correction was needed, thanks to plastic-free equipment, cotton clothing, filtered compressed air for the filters and rigorous glassware cleaning.

The researchers position the protocol as a practical compromise for large-scale monitoring, where throughput and cost matter more than perfect recovery of every high-density polymer. Compared with enzymatic digestion methods that offer high accuracy at the price of weeks of processing, or instrument-heavy approaches requiring expensive equipment and expert operators, the new workflow is fast, inexpensive and compatible with standard microRaman instruments using a 785-nanometer diode laser and the open-source OpenSpecy library for spectral identification. The authors suggest that testing alternative solvents could further improve effectiveness and eco-compatibility, and that a denser flotation salt could extend the method to specific high-density polymers when needed. As European regulators move toward harmonized basin-scale microplastic monitoring under the Water Framework Directive, tools like this one may determine whether such ambitions become routine practice.

Subject of Research: Optimization of sample pretreatment for Raman spectroscopic determination of microplastics in riverbed sediments

Article Title: Advances in the sample pretreatment for determination of microplastics in riverbed sediments by Raman spectroscopy

Article References: Dali, A., Mancini, M., Santianni, D., Solari, L., & Becucci, M. (2026). Advances in the sample pretreatment for determination of microplastics in riverbed sediments by Raman spectroscopy. Environmental Science and Pollution Research. https://doi.org/10.1007/s11356-026-38231-8

Image Credits: AI Generated

DOI: 10.1007/s11356-026-38231-8

Keywords: microplastics, Raman spectroscopy, river sediments, sample pretreatment, Fenton oxidation, cyclohexane extraction, Arno River, density separation, plastic pollution, environmental monitoring, polymer identification, water quality

The study, led by Nazanin Jabbarzadeh and colleagues, stands out in a crowded field for the care it took with quality control. Microplastic research has been criticized for inconsistent methods, missing blanks, and questionable identifications, so the team built their workflow around safeguards. They processed five procedural blanks per batch, exposed blank filters to laboratory air during each filtration session, pre-filtered all reagents through 0.45 micrometer membranes, and wore cotton lab coats and nitrile gloves inside a laminar flow hood to minimize airborne contamination. Procedural blanks averaged just 1.2 particles per filter and airborne blanks 0.6, yielding a detection threshold of 3.6 particles per filter, well below the lowest sample concentration recorded.

Each one-liter milk sample was digested with 30 percent hydrogen peroxide at 60 degrees Celsius for 24 hours, then vacuum-filtered through 1.2 micrometer glass fiber filters. To gauge how much material the method might lose along the way, the researchers ran spike experiments, adding known quantities of standard microplastic particles to ultrapure water and processing them exactly like samples. Recovery rates ranged from 82 percent for polyethylene terephthalate to 96 percent for polyethylene, with an overall mean of 89 percent. The team then corrected reported concentrations using these polymer-specific recovery factors. They are careful to note, however, that because the spikes were performed in water rather than milk, the recovery figures do not amount to a complete matrix-matched validation of the method for the dairy matrix itself.

Microscopic screening under a stereo microscope at 40 and 100 times magnification recorded 551 particles across all 30 samples. Morphologically, the haul was dominated by fragments, which made up 58 percent of the total, followed by fibers at 35 percent and spheres at 7 percent. Fibers were mostly blue or transparent, while fragments displayed a broader palette including black and red. Three trained analysts counted particles independently for each filter, with the mean recorded to limit observer bias, and systematic grid patterns prevented double-counting. Critically, the authors treat visual examination as a screening step only, not as definitive evidence of polymer identity, a distinction that has often been blurred in earlier microplastic studies of food.

To firm up the chemistry, roughly 150 particles, five per sample, were subjected to scanning electron microscopy coupled with energy-dispersive X-ray spectroscopy, or SEM-EDS. Gold-palladium sputter-coated particles were imaged at accelerating voltages of 15 kilovolts and magnifications up to 10,000 times, revealing surface textures and elemental fingerprints. A polyethersulfone particle, for instance, showed a telltale sulfur peak in its EDS spectrum, while polyethylene particles were dominated by carbon. But elemental composition alone cannot settle polymer identity, so the particles were then interrogated with micro-Raman spectroscopy using a 785 nanometer laser, with spectra matched against reference libraries and custom standards of seven common polymers.

The Raman analysis succeeded for 142 of the 150 particles, a confirmation rate of 94.7 percent, with hit quality indices averaging 0.92. Polyethylene emerged as the most abundant polymer at 32 percent of identified particles, followed by polypropylene at 24 percent, PET at 15 percent, polyamide at 12 percent, and polyethersulfone at 9 percent, with the remaining 8 percent unidentified or other. That profile is striking because polyethylene and polypropylene are precisely the polymers most common in plastic milk bottles and food-contact materials, while polyamide and PET are characteristic of Tetra Pak laminates. The authors caution, however, that polymer identity alone cannot pinpoint where each particle originated.

The statistical centerpiece of the study is the packaging comparison. Milk sold in plastic bottles contained an average of 24.6 particles per liter, roughly double the 12.2 particles per liter found in Tetra Pak cartons, a difference that was highly significant. In a multiple linear regression model, packaging type was the only meaningful predictor of contamination, explaining nearly half the variance in concentrations, while fat content and brand showed no significant effects. Adding product volume to the comparison erased the statistical difference, and the authors are explicit that the observational design of the study establishes an association with packaging, not proof that bottles are the direct source of the particles.

Perhaps most consequential for exposure assessment is the size distribution. SEM measurements of 350 particles yielded a mean longest dimension of 68.4 micrometers, with a pronounced peak between 20 and 50 micrometers. Fully 72 percent of measured particles were smaller than 100 micrometers, and 41 percent smaller than 50. This matters because smaller particles are considered more capable of crossing biological barriers and being taken up by cells, although the authors stress that their study measured particle characteristics only and did not investigate uptake, tissue distribution, or any toxicological outcomes. No health risk conclusions can be drawn from these data alone, they emphasize, and milk consumption should not be regarded as a demonstrated health hazard on this evidence.

The study also fills a geographic gap. Most foodborne microplastic data come from Europe, the Americas, and East Asia, while regions such as Iran have been thinly covered despite prior reports of microplastics in milk, milk powder, and infant formula elsewhere. The authors call for larger and more geographically diverse sampling, process-line monitoring inside dairies, controlled packaging-release experiments, and matrix-matched validation of analytical methods, alongside standardized protocols that would make cross-study comparisons meaningful. For regulators, the results argue for routine microplastic surveillance in dairy; for producers, for tighter quality control across processing, storage, and filling. For consumers, the takeaway is sobering but measured: microplastics are now documented in commercial milk, plastic packaging shows the strongest association, and the science of what that means for human health remains an open and urgent question.

Subject of Research: Microplastic contamination and polymer identification in commercial milk using SEM-EDS and micro-Raman spectroscopy

Article Title: Microplastic Contamination in Commercial Milk: Quantification and Polymer Identification Using SEM-EDS and Micro-Raman Spectroscopy

Article References: Microplastic Contamination in Commercial Milk: Quantification and Polymer Identification Using SEM-EDS and Micro-Raman Spectroscopy. (n.d.). https://doi.org/10.1016/j.crfs.2026.101571

Image Credits: AI Generated

DOI: 10.1016/j.crfs.2026.101571

Keywords: microplastics, milk, food safety, polymer identification, Raman spectroscopy, SEM-EDS, packaging, plastic bottles, Tetra Pak, dairy products, dietary exposure, Iran