Comparative physicochemical characterisation of thermally modified wood

Electronic versions

Documents

  • Marcos González Peña

    Research areas

  • PhD, School of Environment, Natural Resources and Geography

Abstract

Small matched samples of beech, Scots pine and Norway spruce woods were heattreated to twenty schedules, at temperatures of 190, 210, 230 and 245????C for heating periods of 0.3, 1, 4, 8 and 16 h. Following thermal modification, physical changes and chemical composition in thermally modified wood (TMW) were analysed and compared to the characteristics of untreated materials. Most of the analysis were carried out using the heat-induced weight loss values (WL, %) as the independ~nt variable to make the comparisons. Average treatment WL values ranged from 0.3 to 27.0, from 0.6 to 21.5, and from 1.0 to 26.7 for beech, pine, and spruce samples respectively. A major finding of this research was that each characteristic changed following the same profile in the three wood species for all properties studied. Save for hardness in beech, the magnitude and rate of change of each property was often found not to be significantly different between species when relative-to-control values in function ofthe WL were compared. The results show that the gravimetric and dimensional changes in the transverse plane are mainly due to chemical breakdown of wood hemicelluloses. On the other hand, the readiness ofTMW to dimensional changes and moisture uptake below the FSP was of a much lower concern than in raw wood. Similarly, several mechanical properties were found to be little affected by heat exposure at treatment levels likely to be achieved industrially. Among these were compression strengths parallel and perpendicular to the axis in the tangential direction, and the axial modulus of elasticity in bending (MOE). Other mechanical strength parameters (e.g. shear strength, hardness and the modulus of rupture in bending) were reduced at almost any given treatment schedule, but the rate of reduction was mild. It is contended that design practices could make up for the strength loss incurred by the treatment for these properties. On the other hand, parameters related to the energy required to produce failure and also impact strength were greatly diminished at low levels of modification. This may restrict the use of TMW where unpredictable sudden loads may occur. Properties at the limit of proportionality were found to be less reduced than the same property at maximum load. Mechanical properties more affected by heat were found to be those more intimately related to the chemical integrity of the wood material and/or the structural configuration of wood polymers in the cell wall substance, whilst compression strength and hardness were more related to wood density. MOE was proposed to be more dependent of cellulose, microfibril angle and wood density, all of these little altered by wood heating. Lastly, a description of the colour changes in the test samples was undertaken using image analysis and qualitatively by infrared spectroscopy. It was determined that colour changes in wood were caused mainly by changes in the lignin. The major input of this work to the state of the art is given by modelling the properties studied using various linear, non-linear and multivariate methods. Properties in TMW were estimated from gravimetric, colour and moisture-related parameters. Solid-state mid-infrared spectra data, treatment parameters and interrelationship between variables were also explored. It is concluded that all changes in small heated samples of wood are amenable to be estimated efficiently in multiple ways, using cost-effective indicators. Remarkably encouraging results for the prediction of physical properties were found for models using colour parameters or by the analysis of the infrared spectra data. Whilst descriptions of the heat-induced changes in wood exist, prediction studies are fairly scarce and not comprehensive. This part of the study is thus very timely, giving the current requirements of quality control and assurance of TMW at the industrial stage. The relevance of these fmdings to property prediction in larger wood members would depend on the property considered; safety factors will probably be needed for some forms ofwood strength in larger heated materials.

Details

Original languageEnglish
Awarding Institution
  • Bangor University
Supervisors/Advisors
  • Martin Breese (Supervisor)
  • Mike Hale (Supervisor)
Thesis sponsors
  • Mexican Council of Science and Technology
Award dateJan 2007