Classification of polyethylene by Raman spectroscopy

Importance of the Topic

Polyethylene (PE) is one of the most widely produced plastics worldwide, with annual output around 80 million tons. Its density, driven by the molecular crystallinity ratio, dictates mechanical strength, chemical resistance, and end-use applications. Traditional density measurements require pure PE samples, making analysis of multilayer films labor-intensive. Raman spectroscopy offers a non-destructive, minimal-preparation route to probe crystallinity and classify PE in situ.

Objectives and Study Overview

• Assess confocal Raman microscopy combined with discriminant analysis to distinguish high-density (HDPE) and low-density polyethylene (LDPE) in both pellets and films.
• Identify spectral markers linked to PE crystallinity and density.
• Validate classification performance using independent samples.

Methodology and Instrumentation

Raman spectra were collected with a Thermo Scientific DXR2 Raman Microscope (532 nm laser, 2 mW; 10× objective; 50 µm slit). Sixteen PE samples (10 pellets, 6 films) of known density were measured at multiple locations, and averaged spectra were generated. Data pretreatment included Norris second derivative and standard normal variate (SNV) to remove fluorescence background and intensity variations. Principal component analysis (PCA) was applied to the C–H stretching (2825–2970 cm⁻¹) and CH₂ bending (1398–1470 cm⁻¹) regions. Mahalanobis-based discriminant analysis in TQ Analyst software classified samples.

Main Results and Discussion

• PCA captured >99% of spectral variance in five factors, with PC2 correlating strongly to density differences.
• Calibration on 12 samples produced clear separation of HDPE and LDPE clusters, with interclass Mahalanobis distances >4.
• Validation on four unknown samples achieved accurate classification (within-class distances <3), for both pellets and films.
• Sample form (pellet vs. film) had negligible impact on the model.

Benefits and Practical Applications

• Non-destructive, in situ analysis of individual PE layers in multilayer films without mechanical separation.
• Rapid measurement (minutes per sample) following method setup.
• Minimal sample preparation reduces labor and time.
• Applicable in QA/QC for packaging, extrusion, and film production.

Future Trends and Opportunities

Next-generation Raman systems (e.g., DXR3) will enhance acquisition speed and sensitivity. Extending multivariate Raman methods can enable classification of copolymers, blends, and contaminants. Integration with automated inline inspection may deliver real-time monitoring in manufacturing.

Conclusion

Confocal Raman microscopy paired with PCA-based discriminant analysis provides a robust, non-destructive method to classify HDPE and LDPE in pellets and films. The approach removes the need for pure sample isolation, offering fast, reliable density assessment for research and industrial quality control.

Instrumentation

• Thermo Scientific DXR2 Raman Microscope (532 nm, 2 mW, 10× objective, 50 µm slit)
• OMNIC software for spectral acquisition
• TQ Analyst software for chemometric modeling

References

• ISO 1183-1/ASTM D792; ISO 1183-2/ASTM D1505; ASTM D4883 – PE density measurement standards.
• Strobl G.R.; Hagedorn W. J. Polym. Sci. B Polym. Phys. 1978, 16, 1181–1193.
• Sato H. et al. J. Appl. Polym. Sci. 2002, 86, 443–448.
• Thermo Scientific TQ Analyst Software, Chemometric Algorithms, 2009.