Microplastics Analysis Using FT-IR, IR Laser and Raman Spectroscopy

Microplastics analysis is the scientific process of detecting, identifying, counting, and characterizing plastic particles smaller than 5 mm in our environment, food, water, or in biological samples. A complete analysis determines not just whether microplastics are present, but how many, what size, what shape, and especially, what polymer type they are.

Visual inspection alone cannot distinguish a synthetic polymer particle from a natural one like cellulose or a mineral fragment. However, this information is very important for identifying, for example, the sources and origin of the particles. Chemical identification (e.g. using spectroscopy) is the only reliable way to confirm a particle is plastic, making it the foundation of any valid result.

Microplastics have been detected in ocean water, deep-sea sediments, Arctic ice, agricultural soil, indoor air, tap water, bottled water, human lung tissue, blood, and placental tissue. 87% of global tap water samples and 93% of tested bottled water brands show contamination. Unfortunately, microplastics are ubiquitous in our modern world, with consequences for humans and the environment that are, in some cases, still unpredictable.

Nanoplastics are plastic particles below 1 µm. Their small size allows them to cross biological barriers that larger particles cannot, including potentially the blood-brain barrier. IR methods cannot reliably detect them; Raman microscopy is the appropriate technique for nanoplastics characterization due to its superior spatial resolution.

The two main approaches are mass-driven analysis and particle-driven analysis. Mass-driven methods like Pyrolysis GC/MS quantify total polymer concentration by mass but destroy the sample, losing all size and count data, making it unable to provide the absolute particle count for the corresponding polymer classes. Particle-driven methods (e.g. using IR or Raman microscopy) identify and characterize every individual particle while keeping the sample intact.

Regulators require particle-driven spectroscopic methods. EU Commission Delegated Decision 2024/1441 and ISO 24187:2024 both specify IR microscopy for particle-based microplastic characterization, defining results in terms of particle counts and size classes — not mass concentration.

The three techniques are FT-IR microscopy, IR Laser Microscopy, and Raman microscopy. All three methods combine a microscopic approach with a spectroscopic technique to measure a molecular vibration spectrum of a particle. Each offers different trade-offs in spatial resolution, throughput, spectral range, and sample compatibility.

Subsampling means analyzing only a small, representative fraction of a sample instead of the entire sample or full filter. In micro-spectroscopic analysis, such as Raman, it is often used to reduce measurement time, especially when particle loads are high. Common subsampling approaches include volumetric aliquots, filter area downscaling, subsectioning, and numerical target subsampling. Each carries specific risks, such as particle loss during transfer, overcrowded filters, radial deposition bias, or statistical error from measuring too few particles.

Subsampling introduces uncertainty because particles are rarely distributed evenly. Clustering, edge effects, filtration artifacts, handling losses, and radial deposition patterns can all cause the measured fraction to differ from the true composition of the full sample. As a result, subsampling can affect estimates of particle number, polymer composition, size distribution, and morphology. Subsampling is therefore a compromise rather than a best practice. Full-filter analysis remains the most reliable approach for quantitative interpretation. It minimizes spatial sampling bias, captures heterogeneity across the filter, and provides the strongest basis for conclusions about particle counts, polymer types, size distributions, and morphology.

FT-IR microscopy uses a Fourier transform IR spectrometer coupled to a microscope to obtain a broadband spectrum of the particle. Some approaches aim at localizing the particles by their visual contrast and take single spectra at the respective positions only. In the case of FT-IR imaging, a variant of FT-IR microscopy, a chemical map of the entire filter is generated, and particles are localized and identified solely based on their spectral signature. FT-IR microscopy reliably detects particles in the micrometer size range, has strong regulatory backing (ISO 24187:2024, EU Decision 2024/1441), and is the recommended starting point for compliance testing, environmental monitoring, and drinking water analysis.

In IR laser microscopy, a tunable IR laser and highly sensitive bolometer detector are used, replacing the classic Globar + interferometer setup. Since the accessible (tunable) wavenumber range is limited in a QCL, the acquired spectra span a reduced spectral range (e.g. 1800 - 950 cm^-1) compared to the FT-IR approach. The use of IR lasers offers considerable advantages when combined with an imaging approach. Because of the high laser power, IR laser imaging (ILIM) illuminates large areas simultaneously, dramatically increasing measurement speed. For example, imaging the entire area of a 25 mm filter is accomplished in approx. 13 minutes. When comparing that to a typical FT-IR imaging measurement of about 2.5 hours (the current gold-standard), it becomes obvious that ILIM is the only technology that can keep up with high-volume screening requirements.

Raman microscopy is used when sub-micron particle detection is required, including nanoplastics research. Its higher spatial resolution allows characterization of particles well below 1 µm, a size range IR methods cannot access. It also excels at identifying inorganic components like fillers within polymer particles. Commonly, Raman analysis workflows rely on prior detection of the particles by optical microscopy and a subsecent selective acquisition of Raman spectra only at those positions.

Sample preparation is the greatest source of variability between laboratories. Inadequate removal of organic and inorganic matrix material leads to false identifications and unreliable results. Contaminants can be tolerated to a certain extent and detected as such by the software. However, if the microplastic particles are obscured or masked by these contaminants, even the software cannot compensate for what arrives on the filter.

Filter choice depends on the measurement technique. Anodisc (aluminum oxide) filters are the industry standard for IR transmission. Silicon membrane filters provide the full mid-IR range and work with Raman. Gold-coated polycarbonate filters are best for Raman and IR transflection. PTFE, metal mesh, and nitrocellulose filters have significant limitations for imaging workflows but could be used for single point measurements with ATR.

Silicon membrane filters are compatible with both IR transmission and Raman measurements. Gold-coated polycarbonate filters work for Raman and IR transreflectance measurements. Anodisc filters fluoresce under laser excitation, show a rough surface at high magnifications, and are therefore not suitable for Raman. 

For imaging approaches, automated software like Bruker's MPID converts millions of raw spectra into a particle list with identity, size, shape, and count for every detected particle. Manual evaluation is not feasible at scale. Machine learning models trained on real-world microplastics spectra significantly outperform traditional library matching, particularly for degraded or contaminated samples.

Environmental particles are degraded by UV exposure, chemical weathering, and contamination. These changes shift spectra away from clean reference standards in ways that simple template matching cannot reliably accommodate. A neural network trained on real-world degraded spectra maintains accuracy where library matching fails.

A confidence score measures how closely the spectral feature vector of an identified particle correlates to a verified reference spectrum for that polymer class. It allows analysts to set a threshold, accepting only high-confidence identifications, to control the trade-off between sensitivity (detecting more particles) and specificity (avoiding misclassification).

The key regulatory documents are EU Directive 2020/2184 (drinking water monitoring mandate), EU Commission Delegated Decision 2024/1441 (specifying IR microscopy as the required method), ISO 24187:2024 (general environmental matrices), and ISO 16094:2025 (water-specific requirements). ASTM WK87463 is under development for North American contexts.

Current regulatory frameworks, including EU Decision 2024/1441 and ISO 24187:2024, specify IR microscopy as the required method for routine microplastic monitoring. Raman microscopy is recognized as appropriate for chemical characterization but is not the primary regulatory requirement for water monitoring workflows.

ISO 24187:2024 covers general principles for microplastic analysis across all environmental matrices (water, sediment, biota). ISO 16094:2025 provides more specific method requirements for water samples only. Laboratories focused on water analysis should reference both.