Wood preservation standards typically specify quality assurance procedures to determine whether wood is adequately treated. As a result there is a need to identify sapwood and heartwood, and measure preservative retention and penetration. For spruce and hem-fir there are no reliable methods to differentiate sapwood and heartwood. For carbon-based preservatives, preservative retention measurement typically requires GC or HPLC analyses; the only methods available to determine penetration involve detecting a surrogate in the formulation rather than the active ingredients. Multivariate models based on near infrared (NIR) spectra have been used to predict a wide range of wood properties over the past 20 years. The present research evaluates the potential use of NIR-based models as quality assurance tools for the wood preservation industry. Models were developed to differentiate hemlock and amabilis fir sapwood and heartwood. Attempts to differentiate spruce sapwood and heartwood were unsuccessful. NIR-based models were also able to differentiate untreated wood from wood treated with DDACarbonate and wood treated with tebuconazole. Models developed to predict DDACarbonate and tebuconazole retention were moderately accurate, but likely not precise enough to replace current quantitative assays. However, the sensitivity to the presence of the actives may be sufficient for estimating preservative penetration. Further work is needed using small probes suitable for scanning increment cores to adapt this technology for industrial use.
In addition to conventional NIR, hyperspectral images were obtained to differentiate untreated wood from DDACarbonate- and tebuconazole-treated wood, but accurate calibrations could not be developed.