A new method for exploring and explaining dimensional stability of commercial and lesser-used wood species from the Congo Basin

One of the main criteria for assessing the potential use of a timber species is the dimensional stability or resistance against swelling/shrinkage. Dimensional changes due to change in moisture content can lead to several defects, e.g. warping, checking, splitting… Thus to support a fit for purpose mentality we need to quantify the dimensional stability of a species. We therefore propose a methodology based on Digital Image Correlation (DIC) to determine the dimensional stability and we will verify whether this methodoloy works for both standardised and non-standardised samples. Specimens were collected from the Tervuren xylarium (Royal Museum for Central Africa, Belgium), which contains over 80000 specimens comprising more than 13000 species (mainly Central Africa), a result of more than one century of collecting efforts, and can be considered as a sleeping beauty in terms of wood technological output. The xylarium offers opportunities to screen a large number of species if a DIC method could be developed, adapted to the various formats of xylarium samples. The axial plane of the specimens was sanded and a 0.5 cm thick section was sawn off to retrieve a sample for DIC. The methodology is based on the calculation of pixel displacements using flatbed scans (SilverFast SE Plus 8 – scanner at 2400 dpi and stored as 48 bit colour images) from the same sample (cross section transversal plane) taken at different humidities (60 % RH – 90 % RH). The results from the DIC method are compared with conventional caliper (Absolute digimatic caliper, Mitutoyo, 0.01 mm) measurements on standardised samples (5.0 x 5.0 x 0.5 cm, 89 samples, 10 species). Wood density (kg/m³) (ovendry/volume) was determined as well on the same specimens using the Archimedes principle. In total 421 non-standardised samples (comprising 53 species) were analysed. The results of the calliper measurements and the DIC method are very similar (linear model: R² adjusted - tangential = 0.96, R² adjusted - radial = 0.94) showing the potential of the digital image correlation method. This method also shows potential towards determining whether the deformation is uniform or rather local. The correlation between average volumetric swelling and average wood density per species was rather weak but significant (r-Pearson = 0.36, p ≤ 0.01). Afzelia species, Pterocarpus tinctorius and Millettia stuhlmannii have a high density and high dimensional stability, thus interesting for exterior joinery applications. On the other side of the spectrum, Antiaris toxicaria and Canarium schweinfurthii have a very low wood density and high volumetric swelling, showing less potential to be used as a high quality timber species. The successful results of this DIC methodology allow for a large screening of non-standardised specimens present in xylaria worldwide