M. Properzi

Wood dowel bonding by high-speed rotation welding

High-speed rotation-induced wood dowel welding, without any adhesive, is shown here to rapidly yield wood joints of considerable strength. The mechanism of mechanically-induced highspeed rotation wood welding is shown here to be due, as already observed in vibration welding, to the temperature-induced softening and flowing of some amorphous, cells-interconnecting polymer material in the structure of wood, mainly lignin, but also of hemicelluloses and consequent high densification of the bonded interface. Wood species, relative diameter differences between the dowel and the receiving hole, and pressing time were shown to be parameters yielding significant strength differences; while relative orientation of the fibre grain of the dowel in relation to the fibre grain of the substrate, relative rate of rotation within a limited range and the use of rough or smooth dowels did not have any significant influence. X-ray microdensitometry and scanning electron microscopy were used to determine the limits of wood dowel welding by high-speed rotation. The type of parameters that had an influence on strength indicated that the strength values obtained, although often rather high, were often due to welding of only a limited part of the dowel to the substrate. This is due to the forcing of the dowel into a truncated conical shape by the pressure of insertion and the consequent disruption of bonding in some areas. Notwithstanding this effect, the welded contact area is sufficient to yield strength results comparable to or even slightly higher than that obtained by PVAc adhesive bonding. The use of dry dowels inserted hot in the substrate after preheating them at high temperature (100°C) yielded consistently better results than that obtained with PVAc gluing.

Parameters influencing wood-dowel welding by high-speed rotation

Oven-dry dowels, insertion of hot dowels, cross-cut dowels, substrate holes of step-decreasing diameter as a function of depth, use of ethylene glycol or other compounds able to decrease the glass transition temperature of wood components have all been shown to contribute to improving weld joint strengths in a variety of less drastic conditions than the 10 mm/8 mm dowel/substrate hole diameter difference. The results show that once the depth of the dowel is much greater than 15 mm, then almost all the conditions used improve the weld strength. This means that the proportion of area welded in relation to the tensile strength of the dowel itself is a determining factor. The greater this area the higher the strength, irrespective of the application conditions used. Thus, over a certain welded area the dowel breaks when tested in tensile, i.e., the joint is stronger than the dowel. Temperatures > 180°C are reached during the quick welding step with the temperature decreasing in less than 1 min to 60–70°C. The same chemical reactions as occurring in vibrational welding have been shown by solid-state 13C-NMR analysis to also occur in dowel rotation welding. In dowel rotation welding the production of carbohydrate-derived furanic aldehydes is higher (a) from the wood material of the substrate in which the hole is pre-drilled rather than from the material of the wood dowel itself, (b) when the weld joint strength is good, and (c) when the rate of dowel insertion is higher.

Properties and set-recovery of surface densified Norway spruce and European beech

The chemistry and wetting behaviour of surface densified wood were investigated using FT-IR spectroscopy and contact angle analyses. Furthermore, set-recovery of the surface under conditions of fluctuating humidity was measured and quantitative microscopy analyses were undertaken. FT-IR indicated that no significant chemical changes took place during the densification process. However, the wettability of the densified surfaces was significantly lower than unmodified surfaces. Following several high humidity-oven dry cycles, it was found that this densification process was almost completely reversible, i.e., there was full set-recovery.

Vibration welding of heat-treated wood

Vibration welding of wood that has been preheated according to an industrial two-step process indicates that such wood can be welded and can yield welded joints of good strength. The joint strength is, however, markedly lower than obtained when welding non-heat-treated timber. In general, weld strength of the timber is poor if welding is done on hydrothermolyzed wood. The strength results are instead much better if welding is done at the end of the complete heat treatment process, i.e., after the dry heat step. The weld lines of heat-treated wood show entangled cells where there is none or very little of the molten matrix intercellular material usually observed in welded timber. Furthermore, in weldlines obtained after hydrothermolysis an increase in rigidity and brittleness of the wood cells is observed. Hence, the wood cells are not entangled at all or very little. Both observations indicate that heat treatment has affected the main melting region of the wood, namely the intercellular material. As most of this material is already either lost or heavily cross-linked during heat treatment, only little of it is now available to melt and bind the wood surfaces during vibrational wood welding.

Solid wood joints by in situ welding of structural wood constituents

Abstract Mechanically-induced wood flow welding, without any adhesive, is here shown to rapidly yield wood joints satisfying the relevant requirements for structural application. The mechanism of mechanically-induced vibrational wood flow welding is shown to be due mostly to the melting and flowing of the amorphous polymer materials interconnecting wood cells, mainly lignin, but also some hemicelluloses. This causes the partial detachment of long wood cells and wood fibres and the formation of an entanglement network in a matrix of melted material which then solidifies. Thus, it forms a wood cell/fibre entanglement network composite having a molten lignin polymer matrix. During the welding period, some of the detached wood fibres no longer held by the interconnecting material are pushed out of the joint as excess fibre. Cross-linking chemical reactions of lignin and of carbohydrate-derived furfural also occur. Their presence has been identified by CP-MAS 13C NMR. These reactions are, however, relatively minor contributors during the very short welding period. Their contribution increases after welding has finished, explaining why relatively longer holding times under pressure after the end of welding contribute strongly to obtaining a good bond.

Comparative creep characteristics of structural glulam wood adhesives

Materials and methods Two commercial phenol-resorcinolformaldehyde (PRF) adhesives and two commercial polyurethane (PU) adhesives, all four of them approved in Germany for exterior grade structural application to glulam and fingerjointing, were tested dynamically by thermomechanical analysis (TMA) on a Mettler apparatus. Six samples of beech wood alone, and of two beech wood plys each 0.6 mm thick were bonded with each adhesive system at ambient temperature (25 C) and under pressure, and then conditioned at ambient temperature and to a constant equilibrium moisture content of 9%. Samples dimensions were of 21 · 6 · 1.2 mm. Triplicate samples were tested in non-isothermal mode between 25 C and 250 C at a heating rate of 10 C/min with a Mettler 40 TMA apparatus in three points bending on a span of 18 mm exercising a dynamic force cycle of 0.1/0.5 N on the specimens with each force cycle of 12 seconds (6 s/6 s). Triplicate samples were also tested in isothermal mode at 40 C, 45 C, 50 C, 100 C and 150 C under identical conditions as above. The classical mechanics relation between force and deflection E 1⁄4 [L/(4bh)][F/(f)] allows the calculation of the Young’s modulus E for each case tested, and this was done to follow the increase of modulus (MOE) as a function of temperature and time. The results are shown in Figs. 1 and 2.

Influence of grain direction in vibrational wood welding

Abstract Wood grain orientation differences in the two surfaces to be bonded yield bondlines of different strength in no-adhesives wood welding. Longitudinal wood grain bonding of tangential and radial wood sections yields an approximately 10% difference in strength results of the joint. Cross-grain (±90°) bonding yields instead a much lower strength result, roughly half that observed for pieces bonded with the grain parallel to each other. These differences can be explained by the very marked effect that homogeneity of fibre orientation is known to have on fibre–matrix composites. Oak yields lower results than beech and maple and is more sensitive to welding conditions. Differences in both anatomical and wood constituent composition can account for this difference in performance. Contrary to the other wood species, oak always presents joint bondlines where little or no increase in density at the interface is noticed. This explains its somewhat lower strength results. This is based on the different mode of bonding predominant in this species, while the other species present two different modes of bonding. Thus, two types of bondlines are observed by scanning electron microscopy (SEM): (i) bondlines where entangled fibre–matrix composites are formed at the interface and (ii) bondlines in which direct welding of the cell walls occurs, just by fused intercellular material or cell surface material. In this latter case the cells remain flat, without an entangled fibre–matrix composite being formed. This is the almost exclusively predominant case for oak. Both cases and even hybrid cases between the two have also been observed in beech.

Wood Welded Connections: Energy Release Rate Measurement

The energy release rates of beech specimens bonded by linear friction welding were determined using double cantilever beam (DCB) tests. The analysis of the results was carried out with the experimental compliance method, which is based on the linear-elastic fracture mechanics. The compliance relation was approximated to a third-order polynomial equation for smoothing and followed by calculation of least squares. This analysis method had not been previously used for glued connections in mode I. It proved to be an ideal method for the results obtained from the study of the energy release rates of the welded joints. The value of the energy release rate G Ic obtained is, on average, 106 J/m2. The variation in the results is less than for most energy release rates of the wood–adhesive couples using beech studied previously. This proved that this method of measurement of the energy release rate adapts well to DCB wood welded specimens. It is the first time that fracture mechanics is applied to the study of linear friction welded wood joints.

Comparative wet wood glueing performance of different types of Glulam wood adhesives

Materials and methods Three different industrial resins were used and compared to bond dry wood at 12% equilibrium moisture content (e.m.c.) and wet wood at the high moisture content of 22% e.m.c. The three wood adhesives used were (i) a commercial polyurethane (PU) adhesive approved for exterior structural grade use in Germany. (ii) A high setting speed phenolresorcinol-formaldehyde (PRF) commercial honeymoon separate application fast-set adhesive approved for structural exterior grade applications in countries such as Australia and South Africa the formulation of which has already been reported (Pizzi et al. 1980, Pizzi and Cameron 1984). (iii) A particularly fast setting rate melamine-urea-formaldehyde (MUF) experimental honeymoon separate application fast-set adhesive of performance already reported (Properzi et al. 2000), the two components of which were composed as follows: Component A: the MUF resin at a pH of approximately 10, solids content of 72%–73%, and viscosity of 2000–2500 mPa.s with no fillers added. Component B: 1.01 parts by weight carboxymethyl cellulose dissolved in 55 parts water, the solution being left to hydrate well for 24 hours and with 27 parts by weight of formic acid solution then added. Beech (Fagus sylvatica) wood strips of dimensions 120 · 25 · 3 mm according to British Standard BS 1204, Part 1, spread with the three adhesives were assembled to have a bonded overlap of 25 · 25 mm, then clamped and left in the clamp at 20 C and the strength increase as a function of time of a set of specimens was tested initially at 15 minutes and then longer intervals to build the curves of strength increase as a function of time shown in the figures. The model fitted to describe the joint strength increment was that of sigmoidal curves of Boltzmann according to the equation y 1⁄4 A2 + (A1 ) A2)/(1 + exp((x ) x0)/dx)). The values obtained for these parameters are reported elsewhere.

Surface modification of wood using friction

The potential of linear vibration friction as an innovative means of producing increases in both surface density and surface hardness was explored. The influence of processing pressure and time on the degree of surface densification, surface hardness and surface elasticity was investigated. It was found that surface hardness (measured as Brinell hardness) was positively correlated with densification ratio. Furthermore, surface elasticity, that is the ability of the surface to recover elastically after indentation during the Brinell hardness test, could be increased by up to 33% depending on the degree of surface densification. The temperature rise due to friction was also studied. During processing, it was found that the temperature rise on both the radial and tangential surfaces was positively correlated with the processing pressure and time.