Thermal modification leads to enhanced dimensional stability among other property changes. There is interest in using mild thermal modifications to add value to joinery products. A mild-modification system for Welsh timber has been developed to improve dimensional stability and working properties of fast grown softwoods such as larch (Larix kaempferi) and Western hemlock (Tsuga heterophylla). There is potential to combine this with lamination to achieve large section pieces suitable for joinery applications. This paper reports a set of tests developed to investigate the strain within glued joints in modified wood under combined wetting, UV and humidity cycling regimes.

The developed method was based on EN 927-6, which is a QUV artificial weathering test typically used to evaluate the performance of coatings on wood. By omitting a mask, and allowing UV exposure over the total surface, the glued joints within the samples were exposed to rapid drying during the UV phases of the cycle and wetting during spray cycles. This places significant strain on any paint films or glue joints, providing a challenging test method, enabling comparison of film performance and bonding. After six days cycling between spray and UV the samples were removed and placed above a water bath to provide elevated humidity on the underside (uncoated face). This had the effect of permitting asymmetric swelling through the thickness of the sample. Samples were removed from the QUV chamber between each one week cycle, to gather data relating to moisture content, degree of splitting, checking or delamination.

The samples were prepared from modified Japanese larch. Pieces were planed and laminated with four lamellae using a polyurethane adhesive. In the final piece the lamellae were aligned radially, tangentially or half-radially or half-tangentially, laid up to minimise and to maximise strain at the glue-line by deliberately favouring matching (e.g. radial to radial) or poor (e.g. tangential to radial) combinations. In addition, combinations of mild and moderate thermal modification were placed in adjacent glue-lines. Samples were tested uncoated, or with clear varnish coating on only the upper face.

Swelling in tangential dimensions of the segments was greater than in the radial dimensions, resulting in visible difference of thickness where radial-tangential bonds had been created. However bonding remained good through the 12 week period. The performance of a high gloss varnish was also monitored, and only minor defects observed in the majority of segments. The observed dimensional changes, moisture uptake and development of splits and checks within the wood will be discussed to illustrate stress development within the laminated samples. The results clearly highlight the potential of this test method as a severe combination of cyclic conditions, with potential for evaluating adhesive performance under asymmetric conditions.