Pressing of wood composite panels at moderate temperature and high moisture content

Pressing of wood composite panels at moderate temperature and high moisture content Robert D. Palardy Bruce A. Haataja Stephen M. Shaler Andrew D. Williams Theodore L. Laufenberg Abstract Aspen flakeboards were pressed at 210 F and 25 percent MC using diphenylmethane diisocyanate (MDI) adhesive catalyzed with a tertiary amine. Three adhesive levels (3.0, 5.5, and 8.0 percent), three pressures (press closing times) (400, 500, and 600 psi), and two catalyst levels (.25% and.50%) were used. Press time was 9.5 to 10.5 minutes. A control group of panels was pressed at 350 F, 7.0 percent MC, and 3.0 percent MDI adhesive for 5 minutes. At the 3.0 percent resin level, the high moisture, moderate temperature panels compared favorably to the control panels in flexural apparent modulus of elasticity (MOE) and thickness stability, but the modulus of rupture (MOR) and internal bond strength (IB) were less than that of the control group. The IB, MOR, and thickness stability of the high moisture, moderate temperature panels improved with increasing adhesive level. The high moisture, moderate temperature panels exhibited a more uniform vertical density profile than did the control panels, but more work is necessary to determine whether the novel pressing conditions reduce internal stresses and improve dimensional stability. Pressing structural flakeboard under high temperatures and low MC is directly responsible for some of the most common panel quality problems. Interior delaminations, or blows, for example, can result from the rapid release of water vapor generated within the mat during pressing. Panel thickness swell and durability may also be worsened by pressing under these conditions due to the accumulation of internal stresses that occurs during a high temperature, low MC press cycle (13). If, however, reconstituted wood panels could be pressed at a lower press temperature and a higher furnish MC, many of these processing problems and quality deficiencies might be eliminated or reduced. Specifically, this pressing technique would: 1. Reduce the amount of energy consumed in the heating of the press and drying of wood flakes. 2. Permit a reduction in the capital cost of new panel plants through a reduction in furnace size and dryer temperature requirements, and elimination of the dryer altogether where air-dry raw material (such as planer shavings) is available. 3. Reduce furnace and dryer emissions. 4. Diminish fire risk through increased wood MC and lower press and dryer temperatures. Conventional structural composite panel pressing methods require a press temperature of 300 to 400 F and a preblending furnish moisture content (MC) of 2 to 8 percent. Meeting these two criteria requires the purchase of expensive drying and press heating systems. Once operating, this equipment consumes substantial amounts of electricity and wood waste-generated heat. Another cost associated with dryers lies in the pollution-control equipment necessary to meet federal and state environmental standards. Finally, the tumbling, conveying, blowing, and storage of wood particles at such a low MC presents a serious fire hazard. The authors are, respectively, Assistant Research Scientist, Inst. of Wood Research, Michigan Tech. Univ. (MTU) Houghton, MI 49931; Senior Research Scientist Inst. of Wood Research, MTU; Assistant Professor, School of Forestry and Wood Prod., MTU; Assistant Research Scientist, Inst. of Wood Research, MTU; and Engineer, USDA Forest Serv., Forest Prod. Lab:, One Gifford Pinchot Dr., Madison, WI 53705. The authors wish to thank Douglas Jurmu, Frank Story, William Yrjana, and Brian Grenley, Laboratory Associates, Inst. of Wood Research., for their assistance with this project. This paper was received for publication in June 1988. Forest Products Research Society 1989. Forest Prod. J. 39(4):27-32. FOREST PRODUCTS JOURNAL Vol. 39, No. 4 27

5. Reduce or eliminate the occurrence of blown panels. 6. Reduce thickness swelling due to moisture absorption by manipulating out-of-press MC to match the expected equilibrium moisture level in service. 7. Reduce panel thickness swell due to springback, if higher wood furnish MC diminishes the buildup of residual compression stress during consolidation of the mat. Background Pressing wood composites at high MC and moderate press temperature undoubtedly produces a range of physical environments within the compressed mat that differs considerably from those produced during conventional pressing. Before considering the potential impact of these novel conditions on consolidation phenomena and resultant composite performance, it is appropriate to review current mat consolidation theory. Then we can speculate on the potential benefits and drawbacks of pressing at moderate temperature and higher wood MC. Under conventional pressing conditions, heat from the press platens rapidly converts moisture near the mat surface to steam. Temperature, MC, and vapor pressure gradients are thus established from the mat surface to its core, and horizontal gradients are generated from the panel center to its edges (1922). The visco-elastic properties of wood particles within the mat depend largely on such localized combinations of temperature and MC (9). Therefore, temperature and MC gradients strongly influence softening and stress relaxation within the mat during pressing (19). Temperature and moisture levels also influence the rate at which adhesive bonds develop (8,9,19). The response of the adhesive to the range of physical environments within the mat will affect the final vertical density profile of the panel, and therefore, final panel properties (19,20,22). Compaction of a particle mat imparts compressive stresses to the individual wood elements. As a result, densification, fractures, and inelastic strains develop within the particles, thereby reducing their inherent strength and stiffness (7,10,18). Though the applied stresses are relieved to some extent by the previously mentioned visco-elastic effects of temperature and moisture on the particles, a complex state of residual stress exists in the finished panel. It is the subsequent moisture-induced release of these residual stresses, in conjunction with microfailures within the wood elements, which accounts for the excessive thickness swell of wood composite panels known as springback (13). The present authors theorize that it is possible to decrease the magnitude of residual stress within finished composite panels-and thus reduce thickness swell-by consolidating the mat at a wood MC significantly higher than the MCs currently used (wood compliance is known to increase with increasing MC). Of course, at conventional pressing temperatures, this would result in intolerably high levels of vapor pressure within the mat. This problem can be overcome by pressing at a lower temperature, but the compliance of wood is also directly related to temperature. Consequently, the changes in wood properties caused by an increase in MC will be offset, to some extent, by the lower temperature, and the net effect on wood compliance will depend on the actual pressing parameters. This interactive relationship between wood compliance, temperature, and MC is currently under investigation by other researchers (9). Pressing reconstituted wood products at high MC and moderate press temperature has not been done in the past because platen temperatures in excess of 300 F were required to cure traditional particleboard adhesives. Diphenylmethane diisocyanate (MDI) adhesive, however, begins to react with aspen wood at 104 to 122 F (21). Isocyanate-water reactions occur rapidly between 230 and 300 F, producing urea- and polyureatype derivatives, which may serve to covalently bridge resin molecules that have reacted with wood (5,6,21). The presence and distribution of moisture affects the products of the MDI-wood reaction, and therefore, curing behavior. In wood composite applications, the effect of mat moisture level on the bonding efficiency of MDI adhesives is unclear. A number of researchers have prepared isocyanate-bonded particleboards at mat MCs ranging from 2 to 25 percent, and platen temperatures of 300 to 400 F. Some investigators (3,4,15) have reported bond strength to be unaffected by changes in MC, while others (11,12) observed a decrease in mechanical properties with increasing mat MC. Plagemann (17) observed an increase in mechanical properties when mat MC was increased from 2 percent to 12 percent. The dimensional stability of isocyanate-bonded flakeboards pressed at conventional platen temperatures has been reported to improve with increasing moisture level (6,12). Objective Given the many potential benefits of pressing wood composites at high mat MC and moderate press temperatures, and the ability of isocyanate adhesives to tolerate high mat moisture levels, the authors decided to investigate the technical feasibility of pressing aspen flakeboards at 210 F and 25 percent mat MC. The objectives of this research were to: 1) compare the properties of panels pressed at 210 F and 25 percent MC to panels pressed at 350 F and 7.0 percent moisture level; and 2) explore the effects of isocyanate resin level, catalyst level, and pressure on the properties of panels pressed at 210 F and 25 percent MC. Procedure Sixteen-inch fresh aspen (Populus spp.) billets were flaked on a 6-foot disc flaker. Target flake thickness was.020 inch and scoring knives were spaced two inches apart. A splitter fan was used to reduce flake width. Flakes were then dried to an MC of 7 ± 1 percent and classified on a Black-Clawson shaker with a 1/4-inch screen to remove fines. Prior to blending, flakes were conditioned to a 25 ± 1 percent moisture level by application of a water spray in a rotary blender followed by sitting overnight in a sealed bag. Amicure SA-102 (Air Products and Chemicals, Inc.), a tertiary amine, was used at two different levels of concentration (0.25% and 0.50%, based on weight of adhesive) to catalyze the reaction of the Mondur E-441 (Mo- 28 APRIL 1989

TABLE 1. Physical characteristics of flakeboards pressed to 1/2 inch thickness at 210 F and 25 percent MC versus panels pressed at 350 F and 7.0 percent MC. a Pressing Adhesive Out-of-press After conditioning b conditions level MC Thickness Density EMC Thickness -------- (%) -------- (in.) (pcf) (%) (in.) 3.0 20.9.647 37.2 9.7.595 (5.2) (3.3) (10.9) (6.1) (2.6) 210 F, 5.5 19.1.591 42.4 9.4.551 25% MC (2.8) (2.6) (4.8) (2.1) (2.7) 8.0 18.4.554 45.3 9.1.523 (4.7) (2.1) (4.2) (13.2) (3.3) 350 F, 3.0 3.7.498 45.7 5.7.504 7.0% MC (5.0) (2.5) (5.3) (11.5) (2.9) a Values for 210 F, 25 percent MC are the average of 18 panels; values for 350 F, 7.0 percent MC are the average of 9 panels. Values in parentheses represent the coefficient of variation (%). b Conditioned at 70 F and 65 percent RH. bay) isocyanate adhesive. The catalyst was first dissolved in acetone, then mixed with the premeasured isocyanate, and applied to the flakes in a rotary blender via an air-spray system. Three levels of adhesive application were used (3.0%, 5.5%, and 8.0%, based on ovendry wood weight). No wax was applied to the flakes. Mat forming was done by hand in an 18- by 18- by 28-inch high deckle box placed over a metal caul sheet. Another caul sheet was placed on top of the formed mat. Flake orientation was random. Mats were pressed without stops in an 18- by 18-inch hydraulic hot-press at three different pressure levels (400,500, and 600 psi). Maximum pressure was maintained until a thickness of 1/2 inch was reached, then pressure was reduced at a rate that maintained that thickness for the duration of the press cycle. Target density was 43 pcf (ovendry wood, not including resin solids) and press temperature was 210 ± 2 F for all panels. Three replications of each board were made for a total of 54 panels. Based on the results of prior experiments conducted at 210 F, it was decided to press each panel for 7 minutes after mat core temperature reached 150 F. This method resulted in a total press time of between 9-1/2 and 10-1/2 minutes for all panels. The authors realize that press time would have to be reduced considerably to make this pressing technique economically feasible, and work aimed at shortening the pressing time is underway. Three conventional panels were also made at each of the three pressure levels. The panels were made in the manner just described, except that press temperature was maintained at 350 ± 4 F and flake MC was 7 ± 1 percent. Three percent adhesive was applied without catalyst. Total press time was 5 minutes. Panels were conditioned at 70 F and 65 percent relative humidity (RH) prior to testing of internal bond, bending strength and stiffness, and thickness swell (immersion method) (1,2). Vertical density gradient was analyzed by the USDA Forest Products Laboratory in Madison, Wis., using a gamma radiation technique (14). Results and discussion The high furnish MC and low platen temperature combined to produce several interesting panel characteristics. Upon press opening, for example, the boards TABLE 2. Mechanical performance of flakeboards pressed to 1/2 inch thickness at 210 F and 25 percent MC versus panels pressed at 350 F and 7.0 percent MC. a Pressing Adhesive conditions level IB MOR MOE M max EI (%) ---- (psi) ---- (ksi) (ft.-lb.) (ft.-lb.) 3.0 35.0 5,323 477.2 912 24,300 (37.8) (11.2) (21.0) (13.9) (29.4) 210 F, 5.5 62.7 7,060 485.8 1,037 19,360 25% MC (35.8) (8.0) (7.6) (10.9) (14.2) 8.0 158.0 8,213 464.2 1,052 15,030 (27.7) (14.5) (9.2) (14.5) (13.1) 350 F. 3.0 97.0 7,370 465.3 888 13,828 7.0% MC (21.8) (15.3) (10.5) (11.1) (14.5) a All calculations are based on actual specimen dimensions after conditioning to equilibrium at 70 F and 65 percent RH. Values for 210 F, 25 percent MC are the averages of 18 panels; values for 350 F, 7.0 percent MC are the average of 9 panels. Values in parentheses represent the coefficient of variation (%). expanded significantly in thickness (Table 1). The amount of expansion ranged from 10 percent to 23 percent, decreasing as adhesive level increased. The thickness expansion in all panels suggested a relaxation of compressive stress upon press opening. Adhesive bonds between flakes inhibit this stress relaxation, as demonstrated by the observed decrease in postpressing thickness expansion with increasing adhesive level. Adhesive level also affected the out-of-press MC of the panels, as shown in Table 1. Boards came out of the press at 20.9, 19.1, and 18.4 average percent MC for the 3.0, 5.5, and 8.0 percent adhesive levels, respectively. A slight drop in panel moisture level as resin level increased was to be expected, since isocyanate is known to react with water in the formation of polyurea bonds (17,21). Panels made under conventional pressing conditions emerged from the press very near the target thickness of 1/2 inch and at an average MC of 3.7 percent. The conventional panels attained an equilibrium MC of 5.7 percent during conditioning at 70 F and 65 percent RH. The high MC panels reached equilibrium at 9 to 10 percent MC. This panel group s moisture loss during conditioning resulted in a 7 to 10 percent reduction in board thickness. The out-of-press thickness expansion and subsequent panel shrinkage during conditioning caused corresponding changes in panel density (Table 1). Average panel densities for the 3.0, 5.5, and 8.0 percent adhesive levels were 37.2, 42.4, and 45.3 pcf, respectively. With panel density as a dependent variable (affected mainly by adhesive dosage), statistical techniques for normalizing data to reflect one mean panel density (covariance) become invalid (16). Therefore, panel performance data is presented with no adjustment for density differences between panel groups, and the results are discussed with reference to these density differences. At a 95 percent confidence level, platen pressure and catalyst level had no significant effect on all panel properties tested. Even though platen pressure affected the press closing time of the high moisture mats, it had no significant effect on vertical density profile, and therefore little effect on the properties measured (Fig. 1). The reason for the lack of catalyst dosage effects is unclear. FOREST PRODUCTS JOURNAL Vol. 39, No 4 29

Figure 1. Representative vertical density profile of a flakeboard pressed at 210 F and 25 percent MC (a) versus a panel pressed at 350 F and 7 percent MC (b). It is possible that the catalyst levels chosen, while typical for many polyurethane applications, were too low to increase reaction rate in the environment of a flakeboard mat. The subject of isocyanate-catalyst-wood reactions is a complicated one, and will be dealt with in greater depth in future publications. The most significant effects on panel properties were caused by adhesive level. The following discussion will focus on these effects and comparisons between high moisture, moderate temperature panels and control panels. Internal bond strength The internal bond (IB) strength of the panels pressed at high MC, moderate temperature, and the 3.0 percent adhesive level was only 36 percent of the conventional panels IB strength (Table 2). The higher density of the conventional panels undoubtedly accounts for some of this difference. It is also possible that the adhesive was not completely cured in the core region of the panels pressed at 210 F. Thermocouples placed in the center of the mats revealed that the highest temperature attained in the core ranged from 198 F to 207 F, compared to 270 F to 291 F for the conventional panels. IB values for the high moisture, moderate temperature panels increased sharply as adhesive level and density increased (Fig. 2). Flexural performance Table 2 shows that the modulus of rupture (MOR) of panels pressed at 25 percent MC, 210 F, and 3.0 percent adhesive was 72 percent of the MOR of the conventional panels. The higher density of the conventional group likely accounts for some of this difference, but vertical density profile may also play a role. Figure 1 illustrates two vertical density profiles; one which is Figure 2. Effect of adhesive level, and the accompanying change in density, on the internal bond strength of flakeboards pressed at 210 F and 25 percent MC. Arrows indicate the performance of panels pressed at 350 F and 7 percent MC. Figure 3. Effect of adhesive level, and the accompanying change in density, on the MOR and MOE of flakeboards pressed at 210 F and 25 percent MC. Arrows indicate the performance of panels pressed at 350 F and 7 percent MC 30 APRIL 1989

TABLE 3. Mechanical and dimensional performance of flakeboards pressed to 1/2 inch thickness at 210 F and 25 percent MC versus panels pressed at 350 F and 7.0 percent MC. a Pressing conditions 210 F, 25% MC Adhesive level Density Thickness b TS c WA d WA;TS (%) (pcf) (in.) ---- (%) ---- 3.0 37.2.595 14.2 62.6 4.41 (10.9) (2.6) (14.8) (14.3) 5.5 42.4.551 6.7 27.6 4.14 (4.8) (2.7) (21.2) (17.7) 8.0 45.3.523 3.1 13.1 4.17 (4.2) (3.3) (20.1) (15.3) 350 F, 3.0 45.7.504 16.1 32.1 2.07 7.0% MC (5.3) (2.9) (14.4) (7.6) a Values for 210 F. 25 percent MC are the averages of 18 panels; values for 350 F, 7.0 percent MC are the average of 9 panels. Values in parentheses represent the coefficient of variation (%). b After conditioning to equilibrium at 70 F and 65 percent RH. c TS = thickness swell. d WA = water absorption. representative of those observed in panels pressed at 25 percent MC and 210 F, and one which was typical of panels pressed under conventional conditions. The high moisture, moderate temperature panels exhibited only a mild density gradient between surface and core, while the gradient of the panels pressed at 350 F and 7.0 percent MC was somewhat more severe. The density profiles reflect the difference in pressing conditions between the two panel groups. The higher platen temperature of the conventional panels produced the characteristic steam shock effect, plasticizing and densifying the material near the panel faces more than the composite core. This effect was diminished in the panels pressed at 210 F and 25 percent MC. MOR increased with increasing adhesive level and density among the high moisture panels, as expected (Fig. 3). While one might have expected the previously mentioned postpressing thickness expansion to affect vertical density profile, no substantial differences were observed within this panel group. The mat MC was apparently high enough to overwhelm the relatively small effects that might have been caused by the independent variables. The modulus of elasticity of the high moisture, moderate temperature panels, on the other hand, was comparable to that of the conventional panels for all adhesive dosages (Fig. 3). The reason for this is unclear. One would normally expect a higher MOR to correspond to a higher MOE. It is particularly interesting that MOE remained constant with increasing adhesive level and density within the high moisture panels. The moment carrying capacity (M max ) and stiffness (EI) of the panels were also determined. The M max and EI of all high moisture, moderate temperature flakeboards exceeded that of the conventional panels (Table 2). M max increased (912 to 1,052 ft.-lb.) with increasing resin content, while flexural stiffness decreased (24,300 to 15,030 ft.-lb.). This behavior results from combining the material properties (MOR and apparent MOE) with the various thicknesses of the flexure specimens. Differences in specimen thickness arose from varying degrees of material expansion upon press opening and the higher equilibrium MC compared to the conventional panel group, as previously discussed. Figure 4. Effect of adhesive level, and the accompanying change in density, on the thickness swell and water absorption of flakeboards pressed at 210 F and 25 percent MC. Arrows indicate the performance of panels pressed at 350 F and 7 percent MC. The influence of specimen geometry is especially pronounced in EI measurements, as the apparent MOE of the panel groups were not significantly different. The acceptability of obtaining good M max performance by use of thicker panels with lower material properties (e.g., MOR) represents a trade-off that must be evaluated by the producer and design engineer. Thickness swell and water absorption Panels pressed at 25 percent MC, 210 F, and 3.0 percent adhesive exhibited a small, but significant (95% confidence level), difference in thickness stability compared to the conventional panels after a 24-hour soak (Table 3), even though they absorbed about twice as much water. It is interesting that the ratio of water absorption to thickness swell for each group of panels pressed at 25 percent MC and 210 F was consistently about twice that of the panels pressed at 350 F and 7 percent MC. This observation supports, though not conclusively, the authors theory that pressing at this high moisture level and moderate temperature reduces moisture induced thickness swell in wood composites. As adhesive level increased, thickness swell decreased sharply among the high moisture, moderate temperature material (Fig. 4). The panels containing 8.0 percent adhesive were exceptionally stable, swelling only 3.1 percent during 24 hours of water immersion. Summary The results of this experiment show that it is possible to make flakeboards of reasonable quality at 210 F and 25 percent MC using an isocyanate adhesive. However, the pressing time required to do so was well beyond the limits of commercial feasibility. The expansion of the panels immediately after pressing may be beneficial in terms of releasing compressive FOREST PRODUCTS JOURNAL Vol. 39, No. 4 31

stress, but pressing parameters such as degree of compression and mat weight must be manipulated to achieve a desired final thickness and density. The potential benefits of pressing wood composites at higher MCs and moderate temperatures, along with the results presented here, lend merit to further study in this area. Current research is aimed at 1) reducing press time to a commercially feasible level through resin catalysis and mat moisture adjustment; 2) evaluating the ability of this pressing technology to improve dimensional stability via the reduction of pressing induced stresses; and 3) investigating the tolerance of this pressing technique to high density hardwoods and species containing high levels of organic volatiles. Literature cited 32 APRIL 1989