Determination of Plasticizer Content in PVC by FT-NIR Spectroscopy

Significance of the topic

Fourier transform near-infrared (FT-NIR) spectroscopy offers a rapid, non-destructive route to quantify additives in polymers. Determining plasticizer content in polyvinyl chloride (PVC) is a routine but traditionally slow and solvent-intensive task (Soxhlet extraction and gravimetry). An FT-NIR approach can provide fast at-line or near-line quality control, minimize solvent use, and deliver near real-time feedback for production control, making it highly relevant for industrial polymer quality assurance and regulatory compliance.

Objectives and overview of the study

The primary objective was to develop and validate FT-NIR methods to quantify dioctyl phthalate (DOP) plasticizer in three PVC sample types: translucent plates, transparent plates, and thin films. The work compared reflectance (integrating sphere) and transmission measurement modes, applied multivariate calibration (Partial Least Squares, PLS) and evaluated prediction performance against known gravimetrically prepared concentrations (5–50% w/w for plates; 9–40% w/w for films).

Methodology

• Sample preparation and sets: Two plate sets (22 samples total) and one set of 11 transparent films. Plates were produced with two commercial PVC grades (Neralit N702 with Naftomix MCS 20 stabilizer; Interlite ZP 7003 stabilizer). Films used Neralit 652 with Landromark LZB 968 stabilizer. DOP levels were controlled during formulation (plates 5–50% DOP; films 9–40% DOP). Sample thicknesses were measured with a digital micrometer (plates ~0.57–1.18 mm; films ~0.234–0.369 mm).
• Spectral acquisition: Thermo Scientific Antaris FT-NIR analyzer; spectral range 12000–3800 cm−1; 90 co-added scans per sample at 4 cm−1 resolution. Plates were measured in diffuse reflectance using an integrating sphere (internal gold flag as background). Films were measured in transmission using a transmission holder with an air background.
• Preprocessing and chemometrics: Multiplicative Signal Correction (MSC) to compensate for scatter and pathlength variations, followed by a Norris second-derivative smoothing/derivative (segment and gap values optimized per data set). Partial Least Squares (PLS) regression models were built using TQ Analyst software. Spectral regions for each calibration were selected algorithmically to maximize sensitivity to DOP-related absorptions while minimizing irrelevant variance.

Instrumentation used

• Thermo Scientific Antaris FT-NIR analyzer
• CaF2 beamsplitter
• InGaAs detector
• Integrating sphere module for diffuse reflectance
• Transmission sample holder for film measurements
• Thermo Scientific TQ Analyst chemometric software (MSC, Norris derivative, PLS)

Main results and discussion

• Spectral behavior: PVC spectra in the NIR are dominated by broad, overlapping overtone and combination bands (C–H vibrations principally). Differences in spectra correlated with DOP concentration and sample morphology (translucent vs. transparent vs. film), requiring separate calibrations for each sample type and measurement geometry.
• PLS model parameters: Optimal models used 3–4 latent factors and targeted spectral windows specific to each sample category. Preprocessing with MSC plus a Norris second derivative reduced baseline and scatter effects and improved linearity.
• Calibration performance (representative metrics):
  • Translucent plates: Calibration correlation ~0.99965, RMSEC ≈ 0.38% DOP; cross-validation correlation ≈ 0.995, RMSECV ≈ 1.49% DOP.
  • Transparent plates: Calibration correlation ~0.99931, RMSEC ≈ 0.59% DOP; cross-validation correlation ≈ 0.996, RMSECV ≈ 1.45% DOP.
  • Films (transmission): Calibration correlation ~0.99983, RMSEC ≈ 0.18% DOP; cross-validation correlation ≈ 0.999, RMSECV ≈ 0.41% DOP.
• Interpretation: Very high calibration correlations and low RMSEC values indicate excellent fit to the training data. Cross-validation errors (RMSECV) are higher for plates than films, likely reflecting greater sample heterogeneity and stronger light scattering effects in thicker or more opaque specimens. Transmission-mode film measurements yielded the most robust and lowest-error model due to reduced scattering and more uniform pathlengths.
• Practical limits: The reported RMSECV values suggest the method can reliably resolve DOP differences on the order of 0.4–1.5 percentage points depending on sample type, which is suitable for many QC applications though trace-level quantification would require further method refinement.

Benefits and practical applications of the method

• Speed and throughput: Spectra acquired in under a minute enable high-throughput QC and near-line monitoring.
• Non-destructive and minimal sample prep: Eliminates extensive solvent extraction and gravimetric steps, reducing consumable use and operator time.
• Versatility: Can analyze samples in reflectance or transmission depending on geometry; useful across a range of product formats (plates, films, foils).
• Process control utility: Rapid feedback supports at-line or near-line control strategies in polymer manufacturing and compound formulation.

Future trends and applications

• Inline/online monitoring: Integration of fiber-coupled probes or belt-mounted NIR systems for continuous monitoring of plasticizer dosing during extrusion or calendaring.
• Calibration transfer and robustness: Development of standardized transfer protocols (instrument standardization, standardization samples, or model updating) to deploy models across instruments and sites.
• Advanced chemometrics: Use of machine learning approaches (support vector machines, neural networks) or Bayesian model updating to improve prediction under varying production conditions and detect outliers or migration effects.
• Broader analyte scope: Extension to other plasticizers, stabilizers, or polymer blends; data fusion with Raman or mid-IR to improve specificity for complex additives.
• Regulatory and sustainability drivers: Reduced solvent usage and faster compliance checks align with green chemistry and regulatory expectations across supply chains.

Conclusion

The study demonstrates that FT-NIR spectroscopy, combined with appropriate preprocessing (MSC, Norris derivative) and PLS calibration, provides a fast, reliable, and non-destructive method to quantify DOP plasticizer in PVC in both reflectance and transmission modes. Transmission measurements on films produced the most precise models, but reflectance-based calibrations for plates also achieved high correlation with acceptable cross-validation errors for industrial QC. The method is well-suited to replace laborious solvent extraction and gravimetric workflows for routine compositional control, improving throughput and sustainability.

Reference

• Thermo Fisher Scientific. Application Note 51593: Determination of Plasticizer Content in PVC by FT-NIR Spectroscopy. 2008.