IN-PLANE DIMENSIONAL STABILITY OF THREE-LAYER ORIENTED STRANDBOARD1 Jong N. Lee? Postdoctoral Researcher and Qinglin Wuy Associate Professor Louisiana Forest Products Laboratory School of Forestry, Wildlife, and Fisheries Louisiana State University Agricultural Center Baton Rouge, LA 70803-6202 (Received November 2000) ABSTRACT In-plane swelling and bending properties of three-layer oriented strandboard (OSB) were investi- gated under the interactive influence of flake alignment level (FAL), flake weight ratio (FWR), resin content (RC), vertical density gradient, and moisture content (MC) levels. Mathematical models based on lamination theories were developed to predict effective modulus (EM), linear expansion (LE), and internal swelling stresses. The model's prediction was compared with actual experimental data. It was shown that the relationship between LE and MC change for OSB was curvilinear with larger expansion rates at lower MC levels. FAL and FWR were two primary variables that significantly affected the magnitudes of LE, modulus of elasticity (MOE), and modulus of rupture (MOR). Increase in RC from 4% to 6% led to little change in all three properties. The model predicted general trends of change in LE, EM, and swelling stresses as a function of FWR at the two alignment and two RC levels. The model's prediction in both EM and LE compared favorably with the experimental data. Prediction of the in-plane swelling stresses showed the effect of the panel MC change and directional dependency. The model provides an analytical tool for opti- rni~ingflake alignment level and panel flake weight ratio to achieve a proper balance between EM and LE for OSB manufacturing. K~jywords: Effective modulus, linear expansion, modeling, panel design, processing variables, struc- tural panel. INTRODUCTION (Bryan 1962). Out-of-plane swelling, known Dimensional stability is an important prop- as thickness swelling (TS), is often accompa- erty in the use of oriented strandboard (OSB). nied by permanent strength loss and some- Special attention to dimensional stability of times product failure. It has been widely stud- wood-based materials has always been de- ied by Jorgensen and Ode11 (1961), Johnson manded since changes in the relative humidity (1964), Halligan (1970), Halligan and Schnie- (RH) of the surrounding atmosphere can affect wind (1974), Lehmann (1978), Geimer (1982), their unit dimensions to a much greater extent Hsu (1987), Liu and McNatt (1991), Xu and than the thermo-effects that control dimen- Winistorfer (1995), and Wu and Suchsland sional changes in nonhygroscopic materials (1997). In-plane swelling, known as linear expan- -1 Member of SWST. sion (LE), can be a very significant factor af- ' This paper (No: 00-22-0688) is published with the ap- proval of the Director of the Louisiapa Agricultural Ex- fecting the state of that exists in the ma- periment Station. terial in both structural and nonstructural uses. Wor~rl<i,rr/ I,hrr E.,e,,<.c,34(1 1. 2002. pp. 77-95 1' 2002 hy thc Society of Wood Sc~rncrand Technology 7 8 WOOD AND FIBER SCIENCE, JANUARY 2002, V. 34(1) The differential swelling behavior of individ- BACKGROUND ual particles or flakes in wood composite pan- In-plane stability of' wood composites els can cause in-plane movements that result in high internal stresses. Thus, dimensional Several experimental studies have been con- stability is also a major factor in the matter of ducted to investigate in-plane stability of durability, since the ultimate failure of an ad- wood composites. Zylkowski (1 986, 1989) hesive bond depends in part on these related carried out comprehensive studies on the di- stresses (Bryan 1962). Where OSB has been mensional stability of structural panels includ- used as an exterior structural material, numer- ing plywood, waferboard, and OSB. LE and ous instances of the so-called "window pane" TS of the panels tested were evaluated. The phenomenon in roof applications have been at- measured swelling data were expressed as the tributed to the disregard of installation clear- percentage of the total dimensional change ance between panels (Xu and Suchsland from the oven-dry to water-soaked condition 1997). There are many factors that affect di- for a given humidity exposure. Results indi- mensional changes of a three-layer OSB cated that the relative LE was nearly constant board. Therefore, a theoretical study should be for all materials and depended upon the panel able to assess the importance of the signifi- MC. The major part of the expansion occurred cance of various factors. at the low MC regions. In an earlier study (Wu 1999), the effects of Using the LE rate, Suchsland and Xu (1 99 1 ) panel processing variables on the in-plane sta- classified wood composites into two main cat- bility behavior of single-layer oriented strand egories. The first category, including medium panels were investigated. It was shown that density fiberboard (MDF), has a swelling rate the shape of the LE-moisture content (MC) that approaches zero at the fiber saturation change curve varied with flake alignment level point (FSP). They called such panels substan- (FAL) and material direction (MD). The var- tially wood-like stable material since the iation was attributed to the difference in the swelling is due mainly to the swelling of the controlling mechanism for LE in various pan- wood cell wall. The second category, includ- els. Total LE from the oven-dry to water-soak ing particleboard, has a swelling rate that condition, modulus of elasticity (MOE), and gradually increases as the materials approach modulus of rupture (MOR) varied significant- the FSP They called such panels substantially ly with flake orientation distribution and den- unstable material since factors other than the sity. LE, MOE, and MOR were correlated with swelling of the wood cell wall contributed to the concentration parameter, density, resin the swelling. One of the factors must be the content (RC), and MC using a power form partial breakdown of adhesive bonds resulting equation. The experimental data form a data- from excessive TS associated with horizontal base of layer properties for modeling three- density variation. This behavior reflects irre- layer, cross-laminated OSB. The current in- versibility of the swelling and of other asso- vestigation dealt with three-layer boards fab- ciated properties. ricated under hot pressing. The objectives of The longitudinal free hygroscopic expan- the study were sion coefficients and elastic recovery after a) To investigate effects of panel processing constrained in-plane swelling were experimen- variables on LE, MOE, and MOR of three- tally determined by Lang and Loferski (1995) layer OSB with density gradient; for commercial plywood and OSB. Their re- b) To develop an analytical model for pre- sults showed that the measured expansions dicting LE, effective modulus (EM), and in- were equal or larger for OSB as compared ternal swelling stresses; and with those of plywood. Approximately 20 per- c) To verify the model by testing three-layer cent of the total hygroscopic expansion can be boards of different construction. considered as elastic deformation for both Lee and Wu-IN-PLANE STABILITY OF STRANDBOARD A) C) FWR = 0 FWR = 1 PSR = 0 PSR = 1 FIG. 1. A schematic of layered structure in OSB showing both single- (a and c) and three-layer (b) panels and three principal material directions (i.e., 1-parallel direction, 2-perpendicular direction, and 3-thickness direction). The rectangles on the graph show the orientation of wood flakes within each board type. panel types. Wu and Suchsland (1996) mea- flakes and the effect of these interactions on sured LE along the two principal directions at dimensional stability and internal stresses. different RH levels for five types of commer- cial OSB. They found that LE changed within Swelling stresses and strains in OSB the hygroscopic range as MC increased in the OSB is manufactured in a layered structure board. Despite differences in wood species by depositing different amounts of wood and manufacturing variables, the broad fea- flakes (based on weight) in the face and core tures of LE were similar among various OSB layers, respectively (Fig. 1). In a given panel, products used. At lower MC levels, LE for all there are three principal directions (i.e., 1- panels occurred at a greater rate and followed parallel or machine direction, 2-perpendicu- the longitudinal swelling behavior of solid lar or cross-machine direction, and 3-thick- wood. At higher MC levels, LE developed at ness direction). Flake weight ratio (FWR), de- a reduced rate and was due mainly to the ef- fined as a ratio of face layer flake weight to fect of transverse swelling of wood. Improve- total panel flake weight, is normally used to ment of FAL and better selection of other pan- describe the layered structure. When FWR = el design variables would reduce this trans- 0 (Fig. la), all flakes are aligned along the verse effect and the overall LE of the panel. perpendicular direction (i.e., single-layer When MC of wood composites changes, in- boards-core). When 0 < FWR < 1 (Fig. 1b), dividual wood components within a panel be- part of the flakes are aligned in the parallel come stressed as a result of the difference be- direction and part are in the perpendicular di- tween actual dimensional change in the board rection (i.e., three-layer boards). When FWR and potential change of the particles or flakes. = 1 (Fig. lc), all flakes are aligned along the Thus, research efforts to improve the stability parallel direction (i.e., single-layer boards- and durability of wood composites could be face). For modeling purpose, panel shelling ra- enhanced by basic information concerning in- tio (PSR), defined as a ratio of face layer teractions between individual particles or thickness to total panel thickness, is more use- WOOD AND FIBER SCIENCE, JANUARY 2002, V.