Mechanical Properties of Individual Southern Pine Fibers. Part I. Determination and Variability of Stress-strain Curves with Respect to Tree Height and Juvenility

This paper is the first in a three-part series investigating the mechanical properties of loblolly pine fibers. This paper outlines the experimental method and subsequent variation of latewood fiber mechanical properties in relation to tree position. Subsequent papers will deal with differences between earlywood and latewood fibers and effect of juvenility and tree height on global fiber properties. In this paper, the mechanical properties were determined on individual wood fiber with a user-built tensile testing apparatus. Cross-sectional areas of post-tested fibers were determined with a confocal scanning laser microscope and used to convert acquired load-elongation curves into stress-strain curves. The modulus of elasticity and ultimate tensile stress of loblolly pine latewood fibers tested in this study ranged from 6.55 to 27.5 GPa and 410 to 1,422 MPa, respectively. Fibers from the juvenile core of the main stem were on the low end of the mechanical property scale, whereas fibers beyond the twentieth growth ring were near the high end of the scale. Coefficient of variation for fiber stiffness and strength averaged around 20 to 25%. The shape of the fiber stress-strain curves is dependent on their growth ring origins: Mature fibers were linear from initial loading until failure, whereas juvenile tibers demonstrated curvilinearity until about 60% of maximum load followed by linear behavior to failure. Keyu~~rkt Modulus of elasticity, ultimate tensile stress, juvenility, confocal scanning laser microscope, cross-sectional area, microfibril angle, INTRODUCTION industry alone are expected to almost double Individual wood fibers are the primary confrom the estimated 1995 level of 191 million stituent of fiberboard and paperboard composmetric tons to a 2010 level of 370 million metites. Global pulp requirements for the paper ric tons (Kaldor 1992). The rapid increase in the manufacture of structural fiberboard panels such as medimn density fiberboard wih com‘IMember of SWST. pound the situation. Substitution of alternate w,,,x/ o,rd I’,lwr SC ,?,I< P. ?-I( I ), X0?. pp i-1 27 0‘) 2 0 0 2 h y Itic Socw~y of WoITtXATLJRI’ R E V I E W Jayne (I 959) was among the lirst of several researchers to develop methodologies for evaluating the mechanical properties of individual wood fibers. Jayne tested fibers to failure in an Instron tensile tester, with fibers mechanically gripped with small jewelers’ vises. Fiber cross-sectional areas of tensile-failed fibers were determined by visual observation under a compound microscope. A modified version of this system was used subsequently by Kellog and Wangaard ( 1964); Tamolang et al. (1967); and by McIntosh and Uhrig (1968). Jayne (1960) reported that a high variation existed in ultimate tensile stress (UTS) and elastic modulus (MOE) both between and within the respective species tested. No values for coefficient of variation or standard deviation were quoted in Jayne’s paper. However, it was speculated that variation may have been attributable not only to fiber geometry and inherent structural differences, but perhaps also to aspects of the testing method employed and specimen selection. Further research by Hartler et al. (1963) concluded that mechanical clamping led to fiber slippage under load. A further major disadvantage of the mechanical restraint method is the unquantified effect of cell-wall compression at the fiber ends, likely contributing to premature fiber failure. Mechanically gripping single fibers often results in more than 50% of the specimens failing at the grips. Hardacker (1963) confirmed this, associating the crushed areas at the fiber tips with low tensile strength values. The problems cited in previous single fiber axial tensile tests spawned the development of a new fiber gripping method. This involved gluing the fibers to various materials, such as paper tabs, and then clamping these paper tabs within the jaws of a tensile tester. Other adhesive gripping methods involved gluing the single fibers to plastic or metal tabs (Klauditz et al. 1947; Van den Akker et al. 1958; McIntosh 1963; Luner et al. 1967; Duncker and Nordman 1965; Leopold and Thorpe 1968). These authors viewed the adhesive method of gripping as a great improvement over mechanical gripping systems. However, problems have also been encountered using this technique. Manipulating the fibers and gluing them is a tedious and time-consuming operation and it can be difficult to obtain an adhesive with suitable properties. The adhesive should not penetrate or flow along the fiber cell wall, but must develop enough adhesion to prevent liber slippage or pull-through. Using the fiber and adhesive system, Hartler et al. (1963) reported fiber misalignment and handling difhculties, both potentially leading to unintentional hber damage. Testing of a 16 WOOD AND FIBER SCIENCE, JANUARY 2002. V. 14(l) misaligned glued fiber would lead to deformation at the grips and to stress concentrations. The result leads to 40% fiber failure rate at the point of gripping (Leopold and McIntosh 1961; Hartler et al. 1963). Ehrnrooth and Kolseth (1984) analyzed hber misalignment and concluded it to be a major contributor to both the development of premature failure at the grips and to the development of unrepresentative load-elongation curves. In a series of experiments conducted at the Pulp and Paper Research Institute of Canada (Page et al. 1972; El-Hosseiny and Page 1975; Kim et al. 1975; Page and El-Hosseiny 1976; Page et al. 1977), fibers were first dried into a flat ribbon shape between a glass slide and a glass micro-cover plate. This prevented the tracheid cell wall from twisting and further facilitated ease of gluing and microanalysis. Improved alignment was noted, and a reduced number of fibers failed prematurely at the end grips. The Fiber Instron developed by Page et al. (1972) also employed a photonic cell to sense fiber elongation, as opposed to gathering data directly from the moving crosshead. This greatly improved the precision in which single fiber strain could be measured. Other unique features of this apparatus included a 16-mm color tine camera attached to a polarizing light microscope. The microscope was mounted in such a fashion as to permit microanalysis of the fiber as the tensile test progressed. This methodology allowed the researchers to establish the relationship between microhbril angle (MFA) and fiber mechanical properties. Kersavage ( 1973) modified the adhesive gripping method in an effort to further reduce fiber misalignment stress concentrations and associated premature cell-wall failure. Drops of epoxy resin were applied near the ends 01 various softwood pulp types. The cured epoxy droplets served as ball joints that were permitted to rotate freely in a ball-and-socket type restraint. Tension was applied to the fiber via the ball-and-socket type assembly. Kersavage (I 973) pointed out that fiber tensile strengths obtained by this method were higher than any comparable published values, suggesting the system causes little strength reducing damage to the individual test specimens. This technique was also adapted to investigate load-carrying capacities of fibers from various locations within a tree (Groom et al. 1995, 1996) as well as various agricultural fibers (Groom et al. 1996). Mott et al. (1996) and Shaler et al. (1996) also used the ball-and-socket type assembly for investigations of failure mechanisms of individual fibers monitored in an environmental scanning electron microscope. Jayne ( 1959) was among the first to develop a database on individual wood fiber mechanical properties. He evaluated the UTS and MOE for 10 softwood species, segregating fiber properties by EW and LW. Results indicated that fibers are generally Hookean in nature, displaying a proportional stress-strain relationship. A limit was found to exist for this region; evidence that wood fibers themselves are viscoelastic in nature. It was conceded that proportional limits varied considerably both between and within each species group. No attempts were made by Jayne (1959) to explain or develop the stress-strain relationships beyond simple speculation, leading Kersavage (I 973) to propose that the curvilinear portions of Jayne’s (1959) load elongation curves (LEC) were due to slippage of the grips. Many authors carried out similar tests very often employing only minor changes in testing procedure in efforts to improve accuracy. Understandably this resulted in similar findings to Jayne’s (1959) tests. However, most investigators realized that the shape of the LEC curve was highly dependent upon test conditions and methods. A more controlled experiment was conducted by Kersavage (1973). The effects of fiber misalignment were considered more carefully, and testing was developed to all but eliminate this problem. Kersavage (1973) was able to report improved accuracy by carefully controlling environmental conditions. Findings suggest that freely dried wood fibers of a FIG. 1. P1iototllicrograph ot’ individual wood libelunder Lensilc load, showing the alignmenr of’ Ihe droplet in the ball-and-socket type assembly. low MFA stressed uniaxially in tension exhibit purely linear LEG, any nonlinearity or curvilinearity (even extreme terminal nonlinearity) being due to the experimental methodology employed. Kersavage (1973) was also able to confirm the findings of Hardacker ( 1963), namely that increasing gauge length has a significant negative effect on ultimate fiber stress and that grip failure was more likely to occur in single delignified specimens than individual lignitied test specimens. It was also reconfirmed that an increased moisture content reduces f iber strength. The average UTS of fibers taken from a Douglas-fir specimen was reported to be from 853-912 MPa and the MOE in the order of 23.5-24.5 GPa. Problems associated with small fiber tensile testing have led to the fact that only a handful of species have been tested in the fashion described (Robalek and Chaturvedi 1988; Jones 1989a; Klungness 1974; McKee 197 I; Oye 1985; Pycraii and Howarth 1980; Van Wyk and Gcrischer 1982). Dif’ficulty in handling such small samples and the time reyuired to conduct microtensile testing have resulted in a lack 01‘ data that relfects only a tiny portion of commercially available species. More impor18 WOOD AND FlBER SCIENCE, JANUARY 2002, V. 34(l) tantly, though, is that these data are most likely skewed, and probably for two reasons. Fiber length limitations precipitated the exclusion of juvenile fibers, thus skewing the data towards cell maturity. In addition, previous researchers were unaware or ignored the aspect of juvenility that affects physical and mechanical properties of wood and wood fiber. EXPERIMENTAL PROCEDURES Muteriul se lec t ion A loblolly pine (Pinus tuedu L.) tree was selected and felled from a conventional plantation stand located at the Crossett Experimental Forest, Crossett, Arkansas. The selected tree, which was 48 years old was straight in form to minimize the presence of reaction wood. The diameter at breast height was 42.2 cm, overall height equaled 30.3 m, and height to live crown was 21.2 m. Immediately upon felling, a disk approximately 2.5 cm thick was removed every 3.05 m in height, starting from the stump and proceeding to a IO-cm top. Several LW slivers were removed from each disk at growth rings 5, 10, 20, 30, 40, and 48. The slivers measured approximately 2 by 2 by 25 mm. The number of growth rings analyzed for each disk varied as a function of tree height, with the uppermost disk consisting of only growth rings 5 and 10. In an attempt to minimize variability, all slivers were removed from the north compass heading. Approximately three slivers per growth ring and tree height were macerated in a solution comprised of I part 30% hydrogen peroxide, 4 parts distilled water, and 5 parts glacial acetic acid. A typical maceration time for LW slivers was approximately 24 h. Macerated fibers were washed several times with distilled water. Dilute fiber slurries were placed between glass slides, thus allowing the fibers to dry without twisting. Dried fibers were placed over a 2.5-mm channel; fiber ends were attached to the channel plate via double-sticky tape, with the center portion of the fibers suspended over the channel. Two epoxy droplets were placed in the center portion of each fiber via forceps with an approximate spacing of 1 mm. An expanded description of the epoxy placement technique can be found in Mott (1995). The epoxy used in this study was Devcon 2-ton, a slow-cure, high-strength, 2-part adhesive. An epoxy:hardener ratio of 56:44 was found to release the most energy during cure as determined with differential scanning calorimetry tests (Mott I995), and thus that ratio was used throughout this study. The epoxy was allowed to cure at 60°C for 24 h followed by a minimum of an additional 24 h at 22°C.