Prediction of volatile organic compound emission from Norway spruce: a chemometric approach combining FT-NIR and TD-GC-MS

Abstract

Continuous and non-invasive analytical methods, such as Fourier transform near-infrared (FT-NIR), are increasingly utilized across various industries, generating substantial data with valuable insights. This study explored the prediction of volatile organic compound (VOC) emission from Norway spruce (Picea abies) building materials using a chemometric approach that combined FT-NIR spectroscopy and gas chromatography-mass spectrometry (GC-MS) analysis. VOC emission from various spruce materials (cross-laminated timber, surface-treated interior spruce panel, and untreated interior spruce panel) was measured using GC-MS, alongside the collection of FT-NIR data from the wood surface. By employing multivariate statistical analysis and predictive modeling techniques, the study found a clear potential of NIR-based models in predicting emission of three key VOCs, $$\alpha$$ α -pinene, hexanal, and benzaldehyde, from spruce building materials. However, the suggested approach showed prediction uncertainty, largely due to a small data set. Refining and validating this chemometric approach necessitate larger data sets and analysis incorporating a broader range of VOCs. For the proposed approach to replace GC-MS in routine applications, further analysis is needed due to the requirement of comprehensive VOC quantification.