| Abstract: |
The increased use of wood as a sustainable building material offers
significant potential for reducing CO₂ emissions in the construction sector.
At the same time, the natural variability of wood products, particularly their
volatile organic compounds (VOCs), poses challenges for indoor air quality.
Monoterpenes such as α-pinene and 3-carene dominate the emissions of pine wood
and are the focus of current research aimed at assessing and reducing VOC
emissions and exposure. The standardized approach to emission analysis, based
on the chamber method according to DIN EN 16516:2017, is time-consuming. This
project report demonstrates that near-infrared spectroscopy (NIR), combined
with multivariate analytical methods, can provide an efficient and precise
alternative. The results show that reliable predictive models for terpene
emissions from pine wood could be developed using NIR technology. For α-pinene
and 3-carene, cross-validated correlation coefficients of R²CV = 0.77 were
achieved, indicating good model quality. The mean deviation in
cross-validation (RMSECV) was 1, 257 μg/m³ for α-pinene and 1, 232 μg/m³ for
3-carene. These models enable a rapid evaluation of product emissions, with
model accuracy performing particularly well at medium and higher
concentrations. However, limitations exist in the lower concentration range,
which restricts the applicability of the method in such cases. The
applicability to wood materials such as strands (long, thin chips) used for
the production of OSB, proved problematic due to transmission effects and high
spectral variability. To make the technology usable for woodbased materials,
advanced calibration approaches are required. Overall, NIR technology
represents a promising complement to existing analytical methods and opens up
new perspectives for the sustainable processing of wood products. |