A FRACTURE MECHANICS STUDY OF THERMALLY MODIFIED BEECH FOR STRUCTURAL APPLICATIONS

Transcription

1 A FRACTURE MECHANICS STUDY OF THERMALLY MODIFIED BEECH FOR STRUCTURAL APPLICATIONS M. Almudena Majano Majano 1, Mark Hughes 2, José L. Fernández-Cabo 1 ABSTRACT: Thermal modification of wood is an established alternative to other preservative treatments that may be aggressive to the environment. When the use of thermally modified wood in structural applications is envisioned, the fracture properties become extremely important due to the significant increase in brittleness. In the work reported herein, the fracture toughness of thermally modified beech (Fagus sylvatica L) under Mode I loading was quantified using Compact-Tension (CT) specimens, loaded under steady-state crack propagation conditions. The influence of three heat treatments levels and three moisture contents, as well as two crack propagation systems (RL and TL) was studied. Complete load-displacement records were analysed and the critical stress intensity factor, K Ic, and specific fracture energy, G f, evaluated. Thermal modification was found to significantly affect the material behaviour; the more severe the thermal treatment, the lower the values of K Ic and G f, with less difference being observed between the most severe treatments. Moisture content was also found to influence fracture toughness, but had a much less significant effect than the heat treatment. KEYWORDS: Compact-tension, fracture toughness, fracture mechanics, Mode I, thermally modified wood 1 INTRODUCTION 123 The thermal modification or heat treatment, of wood is now an established alternative to other preservative treatments that may be aggressive to the environment. Advantages of thermal modification include are an improvement in biological resistance and dimensional stability and a reduction in hygroscopicity [3, 18, 25]. A significant disadvantage is, however, the loss in toughness, accompanied by mass loss, which limits its use in structural applications. Until now, heat treatments have been mainly applied to softwoods and for non-structural purposes. Many studies reporting the mechanical properties of thermally modified wood have appeared in the literature [3, 12, 15, 20, 32], although the findings are difficult to compare due to the variation in treatments and testing conditions. With the abundance of hardwoods such as beech, heat treatment is one way of adding value to this material and such processes have now been commercialised. 1 M. Almudena Majano Majano, ETS. of Architecture, Structural Department, Technical University of Madrid, Avda. Juan de Herrera 4, Madrid, Spain. almudena.majano@upm.es 2 Mark Hughes, Aalto University of Science and Technology, Faculty of Chemistry and Materials, Department of Forest Products Technology, P.O. Box 16400, Aalto, Finland. mark.hughes@tkk.fi Recently, there has been interest in the use of modified wood in true structural applications [26]. When the use of (thermally) modified wood in structural applications is envisioned, however, the fracture properties become extremely important; particularly fracture in tension perpendicular to the grain. Most of the studies employing fracture mechanical approaches to wood have so far been performed on unmodified material. Numerous researches in the field of wood science have applied both linear elastic fracture mechanics (LEFM) and non-linear elastic fracture mechanics (NLEFM). Through the LEFM approach, parameters like the critical stress intensity factor K Ic can be obtained, which characterizes the maximum stress state quantifying the resistance a material possesses to crack initiation. Some violations of the assumptions of LEFM principles [1] are the strain softening process zones that appear around the crack tip [4]. NLEFM concepts are more appropriate in several practical situations [8]. Useful parameters like the specific fracture energy, G f, which is derived from the load-displacement histories, and represents the whole fracture process until complete separation of the specimen including crack initiation and propagation yield, can be obtained. There have been few studies on fracture properties of modified wood reported in the literature to date. Reiterer and Sinn [15] reported the fracture properties of unmodified spruce and spruce modified by thermal treatment and acetylation, finding reductions in

2 toughness of 20% with acetylated wood, and from 50% to 80% for heat treated material. In the work reported in this paper, the fracture properties and fracture toughness of beech (Fagus sylvatica L) thermally modified to three different levels was studied. Tests were carried out using CT tests under Mode I loading and steady-state crack propagation conditions using fracture mechanics principles. The investigations were carried out in the RL (radial-longitudinal) and TL (tangential-longitudinal) crack propagation systems. In order to investigate possible influences of environment conditions on the fracture properties, the behaviour at different moisture contents (MC) was also assessed. Few previous investigations have considered the dependency of fracture behaviour on MC, and those that have, have focused mainly on untreated wood. Most of these studies have employed LEFM principles, and reported values for K Ic [10; 13; 17, 30] with relatively low statistical significance in the range of 12-18% MC and higher values for dry wood. Smith and Chui [21] undertook studies to determine the fracture energy under Mode I loading using bending tests at different MC. The fracture energy was found to increase with decreasing MC from the fibre saturation point to a MC of 18%. Below 18%, any reduction in MC was found to result in a reduction in fracture energy. Reiterer and Tschegg [17] reported an increasing of specific fracture energy with increasing of MC. 2 EXPERIMENTAL 2.1 MATERIALS Beech provided by Mitteramskogler GmbH, Austria, was investigated. The pre-dried (to approx. 8-12% MC) wood boards were modified at heating temperatures of 180º (Mezzo), 200º (Forte) and 230º (Forte exterior) in a dry three-stage heat treatment process (Mitteramskogler 2007). Both the treated and untreated wood samples were obtained from the same log (i.e. twin boards with half thermally treated and the other half left untreated) in order to reduce variability. For the wood fracture test under mode I loading, notched CT specimens were prepared with an initial notch of about 5 mm long, which was then sharpened with a small band saw and finalized with a razor blade cut, about 2 mm long (Fig. 1). In the RL and TL crack propagation systems studied, the first letters, R (radial) and T (tangential) respectively, indicate the direction of the normal to the crack plane. The second letter, L (longitudinal), refers to the direction of crack propagation. Figure 1: Specimen geometry and orientations According to [23], there is not size dependence when the specimen is not too small. In the present study the specimen geometry is too small to obtain size independent fracture values, but all of the specimens had the same dimensions and geometry which allowed the comparison of the experimental values and the determination of the change in fracture properties due to the different heat treatments. Prior to testing the specimens were conditioned at ºC and 33%, 65% and 95% relative humidity (RH) using salt solutions until an equilibrium moisture content (EMC) was reached. The MC of each specimen was determined by oven drying for 24 hours at 103 ºC and remeasuring the weight. MC is expressed as the moisture contained in a sample as a percentage of the oven dry weight. The densities were measured from sets of approximately 30 cubic samples conditioned at ºC and 65% RH. Values varying from 600 to 777 kg/m 3 on untreated beech, and from 472 to 727 kg/m 3 on heat treated beech were obtained. 2.2 PROCEDURE Figure 2: Testing device The Mode I CT tests were carried out using a small stepper motor driven loading device (Fig. 2). The notched specimens were placed on the equipment fixed by two loading pins. The specimens were loaded to failure under displacement control at a speed of 0.4 mm/min. The load was measured with a load cell having maximum load of ±5 kn. Crack mouth opening displacement (CMOD) was recorded using a digital displacement gauge. At the end of the tests, the remaining ligament of the specimen was measured. Complete load-cmod curves were obtained in all the tests. The specific fracture energy (G f ) was calculated from the integrated area under these curves divided by

3 the area of the fracture surface, A, of the specimen according to (Eqn. 1), where F H represents the horizontal splitting force and δ the CMOD: δ max 1 G f = FH ( δ ) dδ A (1) 0 G f is a toughness quantity and characterizes the whole fracture process until complete separation of the specimen including crack initiation and propagation yield. This amount represents the energy needed to produce a unit fracture area including the dissipated energy. The critical stress intensity factor K Ic was calculated according to (Equation 2) [15, 34] -3/2 1/2 K Ic = F Hmax (1- a/h) /(bh ) (2) Where F Hmax is the maximum horizontal force, a is the distance from the line of action of the horizontal force to the notch tip, h is the distance from this line of action to the bottom of the specimen, and b the thickness of the sample (Fig. 1). The expression was applied first to isotropic materials, but recent studies have justified this approach for crack initiation parallel to the grain [19]. 3 RESULTS 3.1 EMC CURVES Prior to the fracture studies, the EMC ranges were plotted, reaching (3.7±1,1)% in 33% RH, (5.4±1.5)% at 65% RH and (15.5±6.7)% at 95% RH for the treated wood; and (7.5±0.3)%, (11.6±0.8)% and (23.4±1.2)% respectively for the untreated material. 3.2 FRACTURE PROPERTIES The load versus CMOD curves were got from the different sets of CT tests. Typical curves on beech treated by Forte treatment compared to those obtained for untreated samples are shown in Fig. 3. The diagram at the top corresponds to the results got in RL propagation system, and the at the bottom to the TL. The curves provide useful information about the fracture behaviour. The solid lines represent the results for heat treated wood, whilst the dotted lines are for the untreated material. The different colours correspond to the different moisture contents. The same scale was chosen for all the diagrams in order to make a direct comparison possible. Figure 3: Representative CMOD versus load (F) curves for untreated (UT) and Forte heat treated (T) beech, in the RL (up) and TL (down) systems, after 33%, 65% and 95% RH conditioning The size and shape of the curves for treated and untreated differ significantly. Differences can be quantified with the following parameters: maximum load, F max, the critical stress intensity factor, K Ic, and the specific fracture energy, G f. F max is a characteristic strength parameter. The F max of treated and untreated twins were totally different, being higher for the untreated raw material, and then decreasing with the severity of thermal treatment. As may be observed (Fig. 3), the load-cmod curves for the heat treated wood show smoother and more uninterrupted behaviour than those exhibited by the unmodified raw material, which display more unstable behaviour following crack initiation. The explanation for this could be related to the complex structure of hardwoods, consisting of a high fraction of radially oriented cells (rays) acting as reinforcement that force the crack to take a more winding trajectory [16]. The maximum load is evident as angular peaks, typical for linear elastic brittle hardwoods [22]. A deviation is noticeable in the case of untreated wood with higher MC which exhibits a more rounded peak at the maximum load. The critical stress intensity factors, K Ic, obtained are shown in Table 1.

4 Table 1: Average values and standard deviations for the critical stress intensity factors, K Ic (MPam 1/2 ), of heat treated and untreated beech, in RL and TL systems for different MC. Mean of 10 measurements. ρ: Density (kg/m 3 ) Heat treated Untreated RH K ρ Ic K Ic K ρ Ic K Ic (%) RL TL RL TL ME FO FE In this case, clear differences in K Ic between the heat treated sets and the unmodified material were found; lower values in the case of the heat treated material indicates that crack initiation occurs more readily in this material. In Table 2 the results for the specific fracture energies G f of beech twins are shown. Table 2: Average values and standard deviations for the specific fracture energies G f (J/m 2 ), of heat treated and untreated beech, in RL and TL systems for different MC. Mean of 10 measurements. ρ: Density (kg/m 3 ). Heat treated Untreated RH G ρ f G f G ρ f G f (%) RL TL RL TL ME FO FE As expected, in both crack propagation systems, G f was found to be higher in the cases of untreated wood than in treated wood. There were no significant differences in G f between the wood subjected to the two severest treatments, FO and FE. Heat treatment produces a degradation of the hemicelluloses, which is the wood component with the highest thermal sensitivity [33]. It leads to a mass loss and plays an important role in strength [6] and is expected to take place at temperatures above 120ºC [7]. It follows that the fracture properties of wood are likely to be affected and this contention is supported by our results. The reduction in the toughness of heat-treated wood can be attributed to its lower plastic ductility. Crack initiation is easier, as is shown by the results of K Ic, and the crack propagation phase consumes less energy and takes place in a more brittle manner in the treated wood. With regard to the crack propagation system, the generally higher fracture values obtained in the RL system show that the stiffness in this orientation is higher than in TL direction. Considering the structure of the hardwood, the higher G f values obtained in the RL orientation for the raw untreated material could be attributed once more to a high volume fraction of radial rays acting as reinforcement in this crack propagation system. According to several authors [2, 5, 16], before macrocrack initiation as well as during the crack propagation phase, the rays carry tensile stresses in the developing process zone around the crack tip (including microcracks and irreversibly deformed regions). Moreover, additional fracture energy is consumed by the rays also act as reinforcement creating fibre bridges behind the crack tip during crack propagation. The ratio of the K Ic values obtained in the RL system between those obtained in TL resulted higher than the ratios in the case of G f. It may indicate a bigger influence of the rays on the crack initiation phase than on the propagation phase. In the case of the treated material, the fracture values reached mostly higher values in RL than in TL system as well, but the differences were slight especially for the severest treatments. The moisture content, MC, is another factor which is known to influence most mechanical properties of wood. Increasing MC leads to a reduction in stiffness and strength. In this study it is clearly shown that MC also affects the fracture properties. Heat treated wood is less hygroscopic than untreated wood. Therefore, it should be mentioned that the MCs obtained for untreated and modified twins by conditioning at the same relative humidity are different (see EMC paragraph), which influences the crack propagation stage. For a low MC a more brittle crack propagation phase can be expected. With increasing MC, the load-cmod curves showed a decrease in the maximum stress state and an increase in the displacements (Fig. 3). For untreated beech, these maximum load values, F max, reported at 7-8% and 11-12% MC (33% and 65% RH) are not significantly different. A more pronounced decrease in load was obtained for a MC of 22-24% (95% RH). With regard to the force versus CMOD curves for the heat treated sets, the differences of F max between every moisture condition were slight, and any reduction in MC leads to a clear trend. The samples subjected to FO and FE treatments behaved very similarly and this is most probably related to the low hygroscopicity of heat treated wood (the average range of MC varied between 3% and 12%). The K Ic values show the same trend as for F max. This result agrees with mostly of the related investigations in this field [10, 13, 17]. G f exhibits important differences to K Ic with regard to the influence of MC. An increase in MC results in an

5 increase in G f, and is in agreement with the results reported by Reiterer and Tschegg [17]. This means that at higher MCs, more energy per unit area is needed to separate a wood sample into two halves. The energy needed to create the process zone for a crack propagation increases due to the ductility, more energy is consumed in fibre bridging behind the crack tip, and also the dissipation energy increase because of the irreversible deformation. The CMOD prior to the crack start is larger with increasing MC, which means that wood can be strained more until crack propagation initiates. Therefore, the increase in G f with increasing MC found from our study could be explained by an increase in ductility that compensates for the reduction in stiffness argued by the findings of K Ic. A deviation from these results was found for untreated wood in the TL orientation, which showed the highest G f at intermediate relative humidity. In the case of heat treated wood, The increase of G f with increasing MC was not found to be in a high range for the two strongest treatments, FO and FE, as happened with the other fracture parameters. ME sets in the TL system at the highest MC, 18%-20%, reached G f close to the values obtained for untreated wood. 4 CONCLUSIONS The influence of three different heat treatments (ME, FO and FE), three moisture contents and two crack propagation system, RL and TL, on the Mode I fracture behaviour of beech has been evaluated using compact tension tests. The load displacement diagrams were used to determine the basic fracture parameters. Heat treated specimens showed a decrease in K Ic and G f in comparison to the unmodified raw material. The more severe the treatment, the lower both K Ic and G f values, as crack initiation was easier and the crack propagation phase consumed less energy and took place in a more brittle way. The stiffness values did not differ significantly between untreated and heat treated wood. Most of the fracture parameters were found to be higher in the RL crack propagation system than in the TL system and may be explained by taking into account the high volume of rays acting like reinforcements in the crack plane. It has strong influence, especially on K Ic. Increasing of moisture contents leads to an increase of ductility, which also produce a significant increase in the specific fracture energy, characterizing the entire fracture process. For the unmodified wood in the TL orientation, the highest G f values have been obtained at intermediate moisture contents and not at the maximum. It must be remembered that for K Ic the opposite trend occurred; a decrease in this parameter being observed with increasing MC. The higher brittleness of the thermally treated wood is expressed in the G f and K Ic parameters. The significant decrease of the fracture properties is to a much smaller extent due to the MC than due to the heat treatment. Thus, the application of a severe treatment penalizes in a sensible way its structural use. These finding are in keeping with other studies regarding the grading of TMT carried out by Widmann et al. [31]. Further studies are currently underway on the application of fracture properties of TMT to find a fracture model for this material. ACKNOWLEDGEMENT The results presented were developed within the IP-SME project HOLIWOOD (Holistic implementation of European thermal treated hard wood in the sector of construction industry and noise protection by sustainable, knowledge-based and value added products). This Project is carried out with the financial support from the European Community within the Sixth Framework Program (NMP2-CT-2005-IP ). This publication reflects the authors view. The European Community is not liable for any use that may be made of the information contained therein. We would like also thank the staff of the Laboratory of Forest Products Technology, TKK, particularly to Pekka Tukiainen, for his support. REFERENCES [1] Aicher S., Reinhardt HW.: Influence of structure size in linear and nonlinear (wood) fracture mechanics. Holz als Roh- und Werkstoff 51: , 1993 [2] Beery WH., Ifju G. et al.: Quantitative Wood Anatomy Relating Anatomy to Transverse Tensile Strength. Wood and Fiber Science 15(4): , [3] Bekhta P., Niemz P.: Effect of High Temperature on the Change in Color, Dimensional Stability and Mechanical Properties of Spruce Wood. Holzforschung 57(5): , [4] Boatright SWJ., Garrett GG.: The effect of microstructure and stress state on the fracture behaviour of wood. Journal of Materials Science 18(7): , [5] Burgert I., Eckstein D.: The tensile strength of isolated wood rays of beech (Fagus sylvatica L.) and its significance for the biomechanics of living trees. Trees 15: , [6] Esteves B., Marques AV. et al.: Influence of steam heating on the properties of pine (Pinus pinaster) and eucalypt (Eucalyptus globulus) wood. Wood Sci Technol 41: , [7] Fengel D., Wegener G.: Wood: Chemistry, Ultrastructure, Reactions, New York, [8] Gustafsson PJ.: A study of strength of notched beams. In: Proceedings, CIB-W18, Meeting 21. Vancouver, [9] Gustafsson PJ.: Fracture mechanics in codes of timber design. In: COST Wood Mechanics Workshop on Fracture Mechanics in Wooden Materials, [10] Kretschmann DE., Green DW.: Modeling moisture content-mechanical property relationships for clear

6 southern pine. Wood and Fiber Science 28(3): , [11] Mitteramskogler GmbH.: Optimised thermal treatment parameters with respect to mechanical properties. Deliverable D8 from the European project NO. NMP2-CT , 45 pp. Project acronym: HOLIWOOD, [12] Poncsák S., Kocaefe D. et al.: Effect of high temperature treatment on the mechanical properties of birch (Betula papyrifera). Wood Sci Technol 40(8): , [13] Prokopski G.: Influence of moisture content on fracture toughness of wood. Int J Fract 79: R73- R77, [14] Reiterer A., Burgert I. et al.: The radial reinforcement of the wood structure and its implication on mechanical and fracture mechanical properties A comparison between two tree species. Journal of Materials Science 37: , [15] Reiterer A., Sinn G.: Fracture behaviour of modified Spruce wood: A study using linear and non linear fracture mechanics. Holzforschung 56: , 2002 [16] Reiterer A., Sinn G. et al.: Fracture characteristics of different wood species under mode I loading perpendicular to the grain. Materials Science and Engineering A332: 29-36, [17] Reiterer A., Tschegg G.: The influence of moisture content on the mode I fracture behaviour of sprucewood. Journal of Materials Science 37: , [18] Rowell RM.: Specialty Treatments. In: Wood Handbook-wood as an Engineering Material. Forest Products Laboratory, Madison , [19] Schachner H., Reiterer A. et al.: Orthotropic fracture toughness of wood. Journal of Materials Science Letters 19: , [20] Shi JL., Kocaefe D. et al.: Mechanical behaviour of Québec wood species heat-treated using ThermoWood process. Holz als Roh- und Werkstoff 65: , [21] Smith I., Chui YH.: Factors affecting mode I fracture energy of plantation-grown red pine. Wood Sci. Technolog. 28: , [22] Stanzl-Tschegg SE.: Microstructure and fracture mechanical response of wood. Int J Fract 139: , [23] Stanzl-Tschegg SE., Tan DM. et al.: New splitting method for wood fracture characterization. Wood Science and Technology 29: 31-50, [24] Stanzl-Tschegg SE., Tschegg EK. et al.: Fracture energy of srpuce wood after different drying procedures. Wood and Fiber Science 26(4): , [25] Tjeerdsma BF., Boonstra M. et al.: Characterisation of thermally modified wood: Molecular reasons for wood performance improvement. Holz als Roh- und Werkstoff 56: , [26] Tjeerdsma BF., Kattenbroek B., Jorissen A.: Acetylated wood in exterior and heavy load-bearing constructions: building of two timber traffic bridges of acetylated Radiata pine. In: Proceedings of the 3 rd European Conference on Wood Modification, Cardiff, UK. Eds. C.A.S. Hill, D. Jones, H. Militz and G.A. Ormondroyd , [27] Tschegg EK., Frühmann K. et al.: Damage and fracture mechanisms during Mode I and III loading of wood. Holzforschung 55: , [28] Tukiainen P., Koponen S.: Fracture behaviour of small wood specimens in RT-direction. In: Proceedings of the 10th World Conference on Timber Engineering. Portland, OR, USA, p. 55, [29] Vasic S., Stanzl-Tschegg S.: Experimental and numerical investigations of wood fracture mechanisms at different humidity levels. Holzforschung 61: , [30] Wang LY., Lu ZY. et al.: Wood fracture pattern during the water adsorption process. Holzforschung 57: , [31] Widmann R., Fischer A. et al.: Mecanically tested parameters and results depending on process parameters. Deliverable D5 from the European project NO. NMP2-CT , 57 pp. Project acronym: HOLIWOOD, [32] Yildiz UC., Yildiz S.: Mechanical and chemical behavior of beech wood modified by heat. Wood and Fiber Science 37(3): , [33] Zaman A., Alen R. et al.: Thermal behavior of scots pine (Pinus Silvestrys) and silver birch (Betula Pendula) at ºC. Wood and Fiber Science 32(2): , [34] Zhao G., Jiao H. et al.: Study on fracture behaviour with wedge splitting test method. In: RILEM Fracture Processes in Concrete, Rock and Ceramics. J. G. M. v. Mier, J. G. Rots and Bakker. E. & F.N. Spon, This paper is based on the manuscript ID WST R1: The Fracture Toughness and Properties of Thermally Modified Beech and Ash at Different Moisture Contents, resubmitted to Wood Science and Technology the 14-Jan-2010 for a second revision and decision