PSI - Issue 10

E. Tsetsekou et al. / Procedia Structural Integrity 10 (2018) 227–234

230

E. Tsetsekou et al. / Structural Integrity Procedia 00 (2018) 000 – 000

4

Fig. 3. Positioning of a pair of fir samples on Instron 3300 machine.

casein performed better than rabbit-skin glue and Vinavil (Fig. 4). Surprisingly, casein's bond strength was higher in the softwood than the hardwood, which is contrasting to the results obtained by all other adhesives tested. In general, softwoods can become less permeable after drying due to pit aspiration (Kamke and Lee (2007)) and thus create weak bonds. Furthermore, bonded wood assemblies typically increase in strength with wood density (Frihar and Hunt (2010)) and thus adhesives are often found to perform better in hardwoods than softwoods (Lavisci et al (2001)). After UV ageing of maple bonded samples, Paraloid B72 demonstrated the highest bond strength followed by fish glue, Vinavil, rabbit-skin glue and lastly casein (Fig.5). It appears that both synthetic adhesives increased their strength compared to un-aged samples (Fig.7). This unusual behaviour is probably due to cross-linking of polymers chains; however similar results have been reported in other studies (Down et al. (1996)). In contrast, natural adhesives showed significant decrease in their bond strength. For fir samples different results were obtained as all adhesives appeared to have lost strength compared to controls after UV radiation (Fig.7), apart from fish glue which showed an unexplainable increase of load. Overall, after UV accelerating ageing Paraloid B72 and fish glue still presented the highest maximum load at break similar to un-aged samples for both species of wood (Fig.5). Following the ageing cycles with RH and T, casein and rabbit-skin glue appeared totally weakened for both species of wood, showing max load at break point lower than ~ 700 N (Fig.6). This behaviour was expected and is aligned with properties reported for these adhesives (Schellmann (2007); Horie (2010); Lambuth (1989)). In contrast, Vinavil appears almost unaffected by this type of ageing, whereas Paraloid B72 has showed to be more influenced as load value at the break point was decreased. Fish glue again presented atypical results as the maximum load value was decreased for maple whereas for fir remained unaffected. It is apparent however that samples exposed to ageing under cycles of RH and T, were generally more affected than the samples exposed to Xenon Arc lamp (Fig.7). This was anticipated, especially for natural adhesives, as fluctuating RH and T can cause changes in their mechanical properties and physical stability (Schellmann (2007); Horie (2010); Lambuth (1989)). It has to be noted that the high standard deviation values obtained, especially for the fish glued samples, are prob ably owed to the variability caused by the way this adhesive is produced and applied. Even though the number of four replicates was recommended by the ASTM D905, it is considered that natural adhesives require higher number of replication in order to minimise this variation and obtain more reliable results. Finally, concerning wood damage during bond break, rabbit-skin glue produced the greatest wood failures of about 3,5-10% of bonded wood surface in fir samples, before and after accelerating ageing with UV and RH-T cycles. Fish glue also demonstrated excessive wood damage in fir samples of about 6-8,5% whereas Paraloid B72 and Vinavil followed. Casein demonstrated no wood failure. Maple samples presented significantly lower percentage of wood failures compared to fir samples. Rabbit-skin glue again demonstrated the highest percentage of wood failure of about 3-4% while fish glue, Paraloid B72 and Vinavil produced failures of approximately 1-3%.