Quantitative Raman analysis of microplastics in water using peak area ratios for concentration determination

Microplastics (MPs), defined as plastic particles ranging from 1 µm to 5 mm, originate from the degradation of larger plastic items and are now detected across marine, freshwater, and river systems, as well as within aquatic organisms and even in human tissues and fluids. Their widespread presence in water bodies, which are closely connected to human consumption pathways, raises concerns about accumulation, circulation, and health effects. Numerous analytical techniques exist for MP detection in water, including optical, thermal, and spectroscopic methods. However, many cannot directly analyze liquid samples and require filtration or separation. FT-IR, while powerful, is strongly affected by water’s broad IR-active peak, causing spectral overlap and reduced sensitivity. Raman spectroscopy experiences significantly less interference from water and provides molecular fingerprinting, enabling both qualitative and quantitative analyses along with particle characterization. Despite its advantages, quantitative Raman analysis of MPs in water remains underexplored, and reliance on absolute peak intensities can be sensitive to measurement conditions. This study addresses this gap by introducing an internal-standard approach that uses the area ratio of MP characteristic peaks to the H2O peak to calibrate and predict concentrations of PE and PVC (0.1–1.0 wt%) in DI water and to validate performance on mixed-polymer samples.

Prior studies have applied diverse methods to detect and quantify MPs. Optical approaches allow rapid, low-cost particle counting but struggle to distinguish MPs from non-plastic particles. Thermal analyses such as TGA-DSC, TGA-MS, and pyrolysis-GC/MS offer sensitive and accurate quantification (e.g., PE and PP calibration via TGA-DSC) but typically target larger particles and are destructive. FT-IR spectroscopy, including FPA-based reflectance imaging, enables rapid, large-area chemical mapping and identification of multiple polymer types on filters; however, it is hindered by strong water absorption for liquid samples. Raman spectroscopy has been widely used for polymer characterization and environmental MPs, with applications including assessment of PE crystallinity/aging, water quality/hardness studies, and quantitative analyses of chemicals using peak intensities/areas. Nevertheless, quantitative Raman determination of MP concentrations directly in aqueous media has been limited, motivating an internal-standard strategy to improve robustness and enable concentration determination.

Sample preparation: PE (40–48 µm) and PVC (40–100 µm) spherical white particles (Sigma-Aldrich) were dispersed in deionized (DI) water to prepare calibration series and validation mixtures. For calibration, separate PE-in-DI and PVC-in-DI series at 0.1–1.0 wt% were prepared. For validation (true concentration), PE and PVC were mixed together in DI across the same 0.1–1.0 wt% range. Concentrations were set by weight in 10 mL DI water (density assumed 1.0). As particles are insoluble, dispersions were stirred at 600 rpm for 30 min at room temperature to ensure dispersion.

Raman measurements: A confocal Raman spectrometer (XperRam C series, Nanobase Inc., Korea) with a 5× objective and a 30 mW, 532 nm laser was used. The scan area was 800 × 800 µm. Each spectrum was acquired for 25 s, and 20 spectra were collected per sample. For plotting and calibration, the 20 spectra were averaged (Gaussian method).

Data analysis and calibration: Characteristic Raman peaks were identified based on literature and measurements. For PE, a strong band at 1295 cm⁻¹ (CH2 twisting) was selected; for PVC, a strong band at 637 cm⁻¹ (C–Cl stretching) was selected. The area ratios of these peaks to the broad H2O peak were computed and plotted versus concentration (0.1–1.0 wt%). Linear fitting (Origin) provided slope, intercept, and coefficient of determination (R²) for each polymer. R² was defined as 1 − Σ(Xi − Yi)²/Σ(Yi − Ȳ)², where Xi is the observed peak area ratio and Yi is the fitted value.

Validation: Mixed PE+PVC samples in DI (0.1–1.0 wt%) were measured, and the resulting peak area ratios (CRaman) were compared with calibration-predicted values (Cref). Prediction performance was quantified by standard error of calibration (SEC = sqrt[Σ(CRaman − Cref)²/n]) and relative SEC (%RSEC, expressed as a percentage of the mean true concentration), enabling assessment of accuracy across the concentration range.

Quality assurance/control: DI water blanks were analyzed to confirm absence of MP contamination prior to mixing. Glassware was used throughout to prevent plastic contamination.

• The peak area ratio approach using H2O as an internal standard produced highly linear calibration curves for MPs in DI water across 0.1–1.0 wt%.
• Reported R² values: PE 0.98537 and PVC 0.99511, indicating strong linear relationships between area ratio and concentration.
• Linear calibration equations (as reported with minor typographical variation in text/figure): for PE, y = 0.25539 − 0.02187; for PVC, y ≈ 0.37139 − 0.0171x (R² = 0.99511).
• In mixed 1.0 wt% PE + 1.0 wt% PVC samples, both polymers’ characteristic peaks were clearly observed in a single Raman spectrum, confirming qualitative identifications within mixtures.
• Validation against calibration yielded low errors: PE SEC = 0.0723214 and %RSEC = 0.48052; PVC SEC = 0.0985576 and %RSEC = 0.52645, supporting accurate concentration predictions from the calibration model.
• Raman spectra showed increasing peak intensities with increasing concentration for both PE and PVC, consistent with quantitative behavior.

Using the H2O band as an internal standard mitigates variability due to measurement conditions that typically affects absolute Raman intensities, enabling reliable conversion of spectral information into concentration estimates. The strong linearity (R² ≥ 0.985) across 0.1–1.0 wt% and low SEC/%RSEC in validation indicate that peak area ratios provide robust quantification of PE and PVC dispersed in water. Detection of both polymers’ characteristic bands in mixed samples demonstrates applicability to realistic scenarios where multiple MP types coexist. By avoiding filtration or sample drying and leveraging Raman’s minimal water interference, the method directly addresses limitations of techniques affected by water or requiring destructive preparation, supporting its relevance for monitoring MPs in aquatic environments, including potential real-time analysis.

The study introduces a Raman spectroscopy-based quantitative method for MPs in water that uses peak area ratios to the H2O band as an internal standard. Calibrations for PE and PVC over 0.1–1.0 wt% show high linearity and are validated on mixed-polymer samples with low prediction errors, demonstrating both quantitative accuracy and qualitative identification within a single measurement. These results suggest feasibility for precise monitoring of MPs in various water bodies, including drinking and seawater. Future work could extend this approach to additional polymer types, broader concentration ranges, complex water matrices (e.g., natural organic matter, salts), and on-site/real-time implementations to support environmental surveillance.