This study investigates the secondary failure of Malaysian Dark Red Meranti (Shorea spp.) and Spruce (Picea abies) finger joints in a glulam beam in a fire test using a bench-scale test set-up. Secondary failure is the occurrence of failure of the bond lines due to fire and the falling off of the outermost tension layers, exposing the uncharred inner layers to a sudden increase of fire intensity. The lack of published work and the difficulties in describing the behaviour of the finger joints after the secondary failure in a full-scale fire test has identified the need for a simple bench-scale method, incorporating the conditions of the standard fire test. This paper focusses on the performance of the finger joints which together with other defects such as knots and splits are generally the weakest component in the glulam beam. The finger joints were bonded with structural adhesives, specifically phenol resorcinol formaldehyde (PRF) and polyurethane (PUR). They were tested in tension to imitate the failure of finger joints on the tension side of a standard fire test of a glulam beam. Constant heat flux was introduced to the finger-jointed specimens to replicate the secondary failure of a glulam beam in the standard fire test. The results of this study indicate a relationship between the charring rate and density of the specimens, with higher density Dark Red Meranti showing lower charring rate compared to the lower density Spruce specimens. Factors such as constant heat flux as opposed to the time-increasing heat flux exposure and specimen size influenced the charring rate of the specimens. The char rate was measured at the early stages of the fire test, which is known to have higher values since the build-up of the charred layers was not sufficiently substantial to protect the inner unburnt wood. Overall, the bench-scale fire test set-up was able to differentiate the fire performance of the adhesives, with PRF showing better fire performance compared to the specimens finger-jointed with PUR adhesive. In addition, tensile tests at ambient temperature showed no significant difference in tensile strength between finger joints bonded with different adhesives for the same wood species. The tensile strengths of the finger joints bonded with different adhesives were influenced by the temperature profile through the joint. The proposed bench-scale fire test was used to compare the quality of the adhesives in a fire situation, specifically with respect to secondary failure. The PRF was selected as the reference adhesive.

There is an increasing use of cross laminated timber (CLT) in the building sector. CLT is a wood panel product made from layers of solid lumber boards. Each layer of boards is oriented perpendicular to adjacent layers and glued on the wide faces of each board. It has been recognised that different adhesive systems have different behaviour in fire; especially that delamination behaviour of CLT can be avoided by choosing a suitable adhesive system. The best method for evaluation of the delamination is a full‐scale fire test, but considering the high costs of such tests, it is of the utmost importance to develop small‐scale methods for evaluating the adhesive bond properties in fire. The intention is that such small‐scale methods should provide the same results as full‐scale tests. A new, smaller scale method for classifying adhesives with respect to  fire properties would also simplify  the planning of  full scale  tests. Previous  tested small‐scale method for evaluation of finger joints is presented in (1).  In this study, a small‐scale fire test methodology for evaluation of CLT adhesive bond performance in  fire  is  introduced  (2).  The  aim  was  to  demonstrate  an  easy  tool  to  distinguish  between  fire resistant adhesive bonds and non‐fire‐resistant bonds, especially with respect to delamination. The cone heater of a cone calorimeter was used to carry out the tests. Cone calorimeter in accordance with ISO 5660 is one of the most widely used bench‐scale instrument in fire research. This small‐scale  device  has  several  advantages  over  larger‐scale  tests  thanks  to  its  fast,  simple  and  cost‐efficient manner to investigate basic material properties.  

A major problem in the manufacture of three-dimensional laminated veneer products (LVP) is damage due to stretching and/or buckling of the veneer. To reduce or eliminate this problem, veneer densification or adding a strengthening layer to the veneer can be an alternative. To study how veneer modification influences the veneer-to-adhesive bond strength, three methods of modification were studied in relation to an unmodified reference veneer: (1) densified veneer, (2) veneer pre-bonded with paper and hot melt adhesive (HMA), (3) veneer pre-bonded with non-woven polypropylene (NW) fabric glued to the veneer (a) with a urea formaldehyde (UF) adhesive, (b) with a mixture of UF and polyvinyl acetate (PVAc) adhesive, and (c) with a PVAc adhesive. Densification, pre-bonding with paper, and NW with UF/PVAc adhesive mixture resulted in no or only a slight decrease in strength of the bond-line compared to the reference. NW glued with UF or PVAc adhesive showed a considerable reduction in the strength of the bond-line. The climatic cycling had no significant influence on the bond strength.

Unseasoned (green) spruce timber side boards of size 25 × 120 × 600 mm were flatwiseglued with a one-component PUR adhesive. Glued pairs of boards were then kiln-dried to 12 % moisture content. A special small-scale specimen for testing the fracture properties of the adhesive bond in Mode I was developed in order to evaluate the adhesive bond properties. The complete force versus deformation curve, including both the ascending and the descending parts, could be obtained with these small-scale specimens, enabling the strength and fracture energy of the bond line to be calculated. In addition, the fractured specimens were examined by scanning electron microscope. Results show that both the tensile strength and the fracture energy of the green glued PUR adhesive bonds were equal to those of the dry glued bonds. The methodology developed and used in the present study gives new possibilities for analysis of the mechanical behaviour of wood adhesive bonds, and particularly of their brittleness and its correlation with the type of fracture path. This is in sharp contrast to the use of standardised test methods (e.g. EN 302, ASTM D905) with specimens having relatively large glued areas. Using such types of specimens, it is not possible to obtain the complete force versus deformation response of the bond. In addition, when using such test methods, failure takes place in the wood or in the fibres near the bond, thus making it impossible to obtain detailed information about the bond line characteristics.

Form defects such as cup, crook, twist and bow, often causes low volumetric and economical yield in dried sideboards of Norway spruce. The high stiffness and density of sideboards, however, make them attractive to use as structural timber. The green gluing process i.e. gluing of unseasoned timber (with the subsequent drying) can make gluing of side boards efficient and can overcome the difficulties in utilization of side boards for structural applications.In present work, a study where computed tomography was used to monitor the drying process of a green glued glulam beam is presented. The beam had a dimension in cross-section of approximately 105×235 mm2 and consisted of eleven sideboards, planed and flat wise glued together with a 1-component polyurethane adhesive. After curing, and prior to drying, the beam was split into two halves, of approximate size of 50×235 mm2.The drying took place in a small drying kiln and computed tomography scanning was done every second hour throughout the drying process to get the density distribution in the beam. When the drying was finished the temperature in the kiln was increased to 103°C and kept for 24 hours, as to get a dry density reference. By use of an algorithm for subtracting the dry density, the moisture evaporation throughout the drying process could be estimated. Despite the harsh drying conditions, with a wet bulb depression of 10°C already from the start of the drying process, no formation of cracks or other quality problems could be seen in the process. Neither could any moisture gradient from the outer to the inner boards be detected.